I welcome our first panel to the first oral evidence in our inquiry into new forms of extremism and radicalisation. We are very pleased to have three eminent academics with us to inform us. Please start from first principles. Do not be worried about levels of knowledge. We want to learn and make sure that we are addressing the right issues in this inquiry. I ask the witnesses to introduce themselves before going into questions. I will start with you, Professor Smith.

I am Laura Smith, a professor of psychology at the University of Bath and director of the Bath Institute for Digital Security and Behaviour, where we have a range of expertise looking at digital security threats, including online extremism.

I am Dr Joe Whittaker, a senior lecturer in criminology at Swansea University and a director at the VOX-Pol Institute. We look at everything related to terrorist and extremist use of the internet and how to counter it.

I am Daniel Allington, a reader in social analytics at King's College London. I am also a fellow of the London Centre for the Study of Contemporary Antisemitism and an associate fellow of the Counter-Extremism Group.

Thank you very much, and we will start our questioning with Ben Maguire.

Thank you, Chair, and thanks to you all for coming in today. Can we start with the definitions and terminology? We see lots of terms thrown around like terrorism, extremism, and radicalisation. What the is relationship between those terms and how do they differ from each other?

There is a huge amount of conceptual ambiguity between these terms. I had the unenviable task of writing a full chapter on this in my PhD and it is an absolute misery. Let us start with terrorism. A lot is often said about issues with defining terrorism, but for the most part both the UK and most other nation states around the world sort of have the essence of what it is. The first general component is violence or the threat of it. The second is that it is ideological in nature, and that will probably be relevant in conversations we have later. Finally, it is intended to cause terror and aims to coerce a population or policymakers as a result. So that is terrorism, and it is something quite narrow. The second question is extremism, where there is a lot of debate and a lot of conceptual ambiguity. While the UK has the definition of terrorism right, we have really struggled with defining extremism over the years and have had a few mishaps. Fundamental British values was clearly a big mistake, if for nothing else than PR reasons. I would follow the lead of scholar J.M. Berger, who talks about extremism being a belief system in which you identify an out-group that is causing some kind of crisis for the group that you feel that you represent, and then you might call it hate or you might label this out-group a potential target for legitimate violence. So you end up basically saying that those are the bad guys and we are kind of giving permission or justifying violence against them. The really important distinction between terrorism and extremism is that terrorism is a behaviour. It is an act of violence or a threat of violence, whereas extremism is usually thought of more as a belief system that may justify and allow the use of violence, but not necessarily. There is always going to be far more people with an extreme mindset than will actually end up acting on it. Radicalisation—the isation suffix—is a process. It is how someone gets from “normal” to either committing violent actions or having an extreme mindset.

I agree with Dr Whittaker. The definition of radicalisation is really important, especially in this context. It is understood as an individual changing to come to support terrorism; a developmental process where somebody's motivation develops and changes. That aspect of motivation is very important because it is not the behaviour, so sometimes a focus on the motivational change will capture a lot of people, whereas behaviour captures only a few. There is a difference in the definitions between the personal change someone might go through with the changing of their attitudes and motivation, and then the behaviour itself which is captured more by the other two definitions.

I do not have a great problem with the fundamental British values, or at least with the content of them. As a branding issue, it was possibly not the best choice of words, but I cannot find anything wrong with the actual values themselves. I want to emphasise the importance of ideology. A crime is not extremist simply by virtue of being extreme. Extremist terrorism and violence is not worse than other forms of violence. What distinguishes it is that it is ideologically motivated and that is why we treat it differently. When dealing with all forms of crime it is important to recognise differences in motivation, methodology and so on, and there are characteristic terrorist methodologies which are there because of the ideological motivations of these kinds of crimes. It is important to emphasise that what we call terrorism is a form of crime and violence; property damage is also included, and I think rightly so. The point is that it is ideologically motivated and the purpose is to intimidate people, lawmakers, public officials, and the general public into doing something differently. For example, it could be as simple as not engaging in particular kinds of business or not dealing with particular people, which is absolutely fundamental to everything we call extremism and terrorism.

Do you think that the current definition of extremism helps or hinders the Government's ability to identify and tackle it?

I am a fan of the current definition, although I would change is one clause slightly. It refers to extremism as an ideology based on hate, and I think that is incorrect because I do not think any extremist would actually see their ideology as being based on hate. Hate is a consequence of extremist ideology. For example, white supremacists would not say that they were motivated by hatred of non-white people; they would say that they are motivated by a wish to keep races separate, let us say. Islamists would say that their ideology is founded on the Sharia; they would not say it is founded on hate. Very few extremists would admit to being motivated by hate. That is the only thing I would want to change.

In the new definition of extremism, No. 3 talks about a permissive environment, and that phrase comes up a few different times as well in the Prevent duty. I absolutely agree that permissive environments are one of the most dangerous aspects that individuals will encounter along the path to radicalisation, but it is ill-defined in such a way that it lacks utility. What is a permissive environment? What are the characteristics? What might it look like and how does it connect with the vulnerabilities and susceptibilities we talk about in the radicalisation process? There is a core aspect of the definition that needs further unpacking.

This moves on from Laura’s point. My only concern is that there is some sense of potential mission creep when you have overly broad definitions. I do not think it is as problematic as it is with terrorism, considering the legislative powers that come with terrorism, but when you have ambiguous or potentially overly broad definitions, I am more worried about collecting things that do not really belong under that name, rather than missing things that may slip through the cracks otherwise.

Why does countering extremism matter, and what are the negative consequences for society beyond the obvious public safety risks?

We do not want Members of Parliament, when they are about to make a vote, to be worrying about their personal safety if they vote a particular way. We do not want members of the public to have to make decisions based on the worry that they may be attacked if they do a particular thing. We do not want companies to make investments based on the perception that there might be an attack on a facility if it is engaged in certain contracts. We do not want people to be looking over their shoulders and worrying about what extremists want them to do.

I think the “so what?” question stems from my last answer. We have extremely robust laws when it comes to terrorism, which have their own set of critiques and so on, but the question was about counter-extremism. If you do not have a strategy and things in the counter-extremism space, you can end up with mission creep, where way too much stuff falls under the counter-terrorism guise, which will lead to a lot of arrests, securitisation, and bad outcomes. Having something else that exists underneath will have a host of hopefully beneficial societal outcomes.

The definition of extremism here sums up why we needed it. It is to prevent harm in lots of different arenas. Extremism, being a broad definition, encompasses a range of harms that we want to prevent, particularly with young people. With the new forms of extremism we are seeing, it is really important that we can capture those and mitigate them.

Dr Whittaker, I have a follow-up on your point about mission creep. Is there any evidence that labelling as extremist organisations or groups that are on the border of maybe only following some definitions of extremism changes the behaviour in those groups? Does calling them extremist drive people to become more extremist?

Absolutely. We are talking about the online domain here and there is a lot of evidence that the way that most social media platforms, particularly the very big ones, are acting now is they are delineating mostly between legal and illegal speech. Typically, those platforms are very good at removing anything that you give the terrorist label to, so they are modifying behaviour—on these bigger platforms at least—to evade those content moderation points. However, wider analysis of the whole ecosystem suggests that the same groups are still inhabiting the platforms that have a high level of operational security. So what you end up getting is the big platforms, which are typically more compliant with national law, trying to stay on the right side of that line and then outlinking to Telegram, Rocket.Chat, or other platforms where the behaviour is clearly illegal and might have branded terrorist content or crisis footage or things like that on it. So it is modifying behaviour, but it does not mean that they are posting in a less extreme manner as a result.

I would like to come back on that because we hear and think about mission creep in this arena, but we have obviously heard critiques of some systems and processes, such as the Prevent system, from William Shawcross and others, where the criticism is that actually definitions and the pool are too narrow in terms of capturing some threats and risks. Do you think that is a fair assessment, or do you think the issue of mission creep is far more significant? Are we capturing too many people within those parameters, or too few?

Prevent, yes, absolutely. There are a lot of different ways to think about how our national counter-terrorism, counter-extremism and violence fixation programmes should work. I am of the view that Prevent should be very narrowly focused on counter-terrorism, and we should think about a different sort of strategy. For the almost exclusively young boys who are caught up in it and are in Prevent, I do not think it is best for about nine out of 10 of them to be going through a counter-terrorism programme as a result.

We will come back to Prevent, but do you think that the balance is struck correctly? I suppose I was trying to draw out of that whether you think that varies across different types of groups and threats.

You mean different ideologies?

Yes.

There is certainly a different way in which different groups are treated. Again, Shawcross made that point that there is this double standard between Prevent and the far-right being treated as this wider thing, whereas for jihadist Islamists being very narrowly focused on proscription. There is a whole host of rule of law issues as to why you might not like that, but it is probably not hugely effective either.

I was interested to draw some of that.

Coming back to the earlier point on the breadth of the definition and the implications of changing the definition, we should be mindful of making something classed as extremist that was not previously. It can have an impact of polarising those groups and making them perhaps more anti-establishment, and then they may then advocate more extreme tactics and actions. There is a balance in protecting people's right to protest while also protecting the public from harm of extremist actions. We should be mindful of the polarisation aspect.

I suppose that balance is between the polarisation and ensuring that we are not missing people at an early stage who are a risk to national security.

A very clear message that came through from William Shawcross's report was that the net is being cast far too widely with regard to what was called mixed, unstable and unclear ideology. The report made clear that more than a majority of Prevent referrals were for mixed, unstable and unclear ideologies when there is no evidence of any terror threat coming from that direction. Clearly something has gone very wrong there and it is a waste of resources. Many of the people may have mental health issues or need other forms of support, but branding them extremists is completely unhelpful. It means that the minority of Channel cases which are genuinely of some sort of extremism concern are getting less of the resource devoted to them. Shawcross also brought out the problem with the far-right being defined too broadly and Islamism being defined too narrowly. There is a particularly great problem within Prevent and other aspects of our counter-extremism policy which is that an ideology known as cultural nationalism has been mistakenly identified as extreme right. The research base that that categorisation was based on absolutely does not support any such categorisation. Cultural nationalism is a far-right movement, but it is not inherently an extremist movement. The mistaken classification of that, which arose from a misreading of the academic literature, has probably led to excessive numbers of referrals in that direction.

What do you all think about extremist behaviours changing? Do you think it is being driven by social media and the internet on the normal channels, or do you think that is changing as well? I know you alluded to it in a previous question.

I think it is changing. We have been investigating a range of different groups and how they use digital communication technologies. It is not so much that organised groups are using the regular mainstream platforms that we have all heard of; we are talking about messaging platforms that are decentralised and harder to access. These are very young people and children using messaging platforms to exchange content, communicate with each other and find other channels to communicate with more extreme content. It is moving away from this more centralised base where you find people, for example, talking about a single ideology into these messaging platforms where they pick and choose different ideological elements to provide this kind of structure and aesthetic for extremist actions.

I completely agree. We are at a really difficult juncture when it comes to trying to regulate these spaces. We had a point probably about 10 years ago when platforms started to be a lot more serious when it came to explicitly material to do with the so-called Islamic State and realised that it was not really tenable to have them on the Facebooks, Twitters and so on, so they ended up getting booted off. Like any deviant actor, they were relatively smart about the next set of places that they inhabited. We are now trying to think about how we put our legislation into action, not just the UK but the EU, Australia and Canada. Specifically terrorist groups in these online spaces are probably more resilient than they have ever been because a lot of these platforms sit outside our jurisdiction. We have seen in the last few months several platforms that host terrorist or extremist content saying, “We do not like the Online Safety Act, so we are going to geo-block British users.” In principle, that sounds like a very nice solution, and we can say, “You either obey our rules or you do not serve British customers.” The only problem is that it takes someone about three and a half minutes to instal a VPN and get around that—even my seven-year-old kid could do that—so it is not actually an effective way of stopping them. We have a problem that we have never had before because the speed of technological change has advanced so rapidly to these decentralised end-to-end encrypted platforms that we are lagging five-plus years behind.

I completely agree; there is a problem. We have seen a number of high-profile cases where people were prosecuted for extremist content posted on the platform formerly known as Twitter, X.com. The thing is that the serious extremists—the people who are really serious about radicalising people, about organising extremist activity—are not stupid enough to post there, or at least not stupid enough to post that kind of content there. It is easily accessible and very visible which is why these people get prosecuted, but they are often not the people who really ought to be the focus of police attention, because they have learned to hide themselves.

What kind of things do you think an extremist might do that do not reach the threshold of terrorism? I am quite worried about what you have said already; is it a risk to national security?

I can talk about the psychological processes involved. The creation of the permissive environment is sub-threshold; people are not necessarily going to be arrested for mobilising others, encouraging violence, or encouraging politicised views around ideology. People can easily create a permissive environment; we have influencers who are posting things that speak to existing sociopolitical grievances and narratives in society who are quite deliberately doing that to stir up those grievances. That is part of the permissive environment aspect; it is the legal but harmful content that we see from these platforms.

Just to add to that, we are currently undertaking a meta review looking at social media recommendation algorithms. While there is not a lot of evidence to suggest that there is a lot of illegal content being amplified on those algorithms, perhaps because platforms have become a bit more adept at removing the illegal stuff, there is quite a strong body of evidence to suggest that that legal but harmful content is being amplified. That content could be borderline extremist content, health misinformation, disinformation, or body image stuff for kids. This is all the bread and butter of these platforms' business models.

I very much agree with what has just been said, but the problem of a permissive environment extends beyond the online space. For example, having an extremist presence on the streets creates a legitimising environment for other forms of extremism in sometimes an even more dramatic way. Of course, the presence of these movements on the streets then becomes distributed online, and it serves to create the impression of a majority that does not necessarily exist in reality but might serve to encourage people to feel permitted to engage in more extreme behaviours.

Do you think antisemitism, misogyny and conspiracy theories are new forms of extremism? Are they, or should they be, a huge concern?

Those three things have been part of every extreme movement for the last at least 100 years—you could not have picked three better things. Be it jihadism or far-right, it is always the Jews' fault, and there is always some conspiracy—usually it is them controlling the world—and they treat women poorly. We have clearly seen in the last five or so years a lot more movements and, in some instances, acts of political violence of which misogyny is the driving force in and of itself, but they are by no means new. All those things have been central to extreme movements for a very long time.

Conspiracy narratives are very important. One of the big drivers of radicalisation is grievance narratives, and many of these take the form of conspiracies. It is the idea that we cannot have nice things because these terrible people have manipulated everything. Who are those terrible people? An awful lot of the time, it is the Jews who are blamed. As we know, the Jews are the sort of big bogeyman of the far-right. It is also the same for Islamists. I have just published a report with the Counter Extremism Group, which focused on the role of antisemitism within Islamism. Antisemitism is as deeply embedded into Islamism as it is into Nazism to the extent that the statistical analysis that I have carried out and included in that report—Islamist Antisemitism—strongly suggests that antisemitism should be considered a risk factor for radicalisation. I would strongly recommend that antisemitism be seen as a diagnostic marker for extremism. Conspiracy thinking is absolutely fundamental to all the extremist movements. It provides a way for extremists to see themselves as good. This is why I do not like that word being based on hate. It is very important to understand that extremists do not understand themselves that way; they understand themselves as saving the world. What are they saving it from with these terrible acts? They are saving it from an imaginary threat, but one that they very much believe in, which is articulated through the language of conspiracy theory. As my colleague Dr Whittaker has emphasised, it is very often the Jews who are identified as the primary conspirators and who are therefore the most likely to be targeted by extremists, especially the extreme right and Islamists.

To organise all those threads, there are groups that are commonly framed as threats—women and Jews, for example—and conspiracies are the organising narrative around that. Misogynistic views absolutely permeate every group I have looked at in the online space. More concerningly, from my perspective this new young people extremist threat that is emerging, very much focusing on young females as the target. Children are the perpetrators, focusing on young females as their victims.

I have a follow-up question looking into conspiracy theories specifically. Conspiracy theories are nothing new. We had people believing that Elvis was still alive way before the internet. Is there something categorically different about conspiracy theories in the age of the internet, or is it just that we are a bit more exposed to everyone's batty views? If there is something categorically different in the internet age, is that dangerous in and of itself?

Conspiracies are a narrative, and just like any narrative, they capture because they resonate with genuine grievances, for example, economic grievances or political grievances, and they are a way of framing the situation for people that is really seductive. Conspiracies are not new, but the internet means that they take off because they are resonant with people; they are persuasive. We have talked about the business models already, but any content that is popular will just keep getting propagated.

A good thought experiment is that 50 years ago if you thought the moon landing was fake, you might have to go quite a long way to find another person who also thought that; now it is three clicks away on your phone. In that sense, the most basic format of the internet is that you can instantaneously and very cheaply find someone on the other side of the world who has the same views as you. In terms of the actual historical question, it is difficult. About once a year the main polling companies will look at conspiracy theories. I do not know how far back it goes or how much it has changed, but most of the kind of main conspiracy theories that you would think of usually have up to about 10% of public support, which is very scary, but tells you something about your question, because it is a much smaller proportion of people who will actually go and commit political violence as a result. I am hesitant to even use the word indicator, while it could be something that is part of the process, in and of itself, it is never going to predict violence.

I agree with everything that has been said so far. I would not see the internet as having changed conspiracy beliefs in any fundamental way; it just puts believers in touch with each other and allows them to disseminate their views in ways that they would not have before. It would be very difficult to get an article in The Times promoting some crazy new conspiracy theory, but anybody can put it online, and then if it gets clicks it gets more clicks because that is how the algorithms work. There have been some changes. I would say one of the main drivers around conspiracy culture currently is the Israel-Palestine conflict. This is understood in conspiratorial ways by quite a large proportion of the population, and those narratives are pushed by extremists.

Dr Allington, I want to go back to a couple of things we reflected on in an earlier exchange. In my understanding, we were talking about the breadth of definitions and you said that, in your view, far-right cultural nationalism is not necessarily extremism, even though perhaps it has been captured by the breadth of some definitions we use. You have reflected that antisemitism is perhaps an important marker in radicalisation and extremism across the board. Obviously, antisemitism is a feature of movements of the left and the right, but in my own mind at least, perhaps there has at times been a feature of far-right cultural nationalism. Would you consider that antisemitism in that way is the point where you move from far-right cultural nationalism to far-right extremism? Is that how you were characterising that?

It was not what I meant to say, but funnily enough it does work like that. We can trace this back to the 1980s to some work that Professor Michael Billig, a very eminent social psychologist, did looking at publications by the National Front, which at that time was the main extreme right-wing movement. He found that there was a division in the National Front between a larger populist wing, and a sort of paramilitary wing which was smaller and much more committed to violence and terrorism. One of the things that divided them at that time was antisemitism. The paramilitary part of the National Front milieu was intensely antisemitic; the populist part was more focused on anti-immigration narratives, anti-black racism, that sort of thing, and much less committed to violence. So, it has actually worked out that way.

Could one not lead to the other in terms of radicalisation? Is that not the path we need to look at?

Yes, of course, but people can move in all sorts of directions. People can move towards extreme views and then move away from them later.

Interesting. I know we are slightly pressed for time, but I just want to explore once again the role of some online spaces and emerging technologies in drawing young people into extremism, which is obviously an area of huge concern across the board as well as for the Committee. What trends do you see in the way that online spaces are actively being used to spread and reinforce new forms of extremism?

As I mentioned earlier, these more encrypted messaging platforms are becoming more common than what we might consider mainstream platforms that the general public use. There is more private messaging and more channels for communication where maybe 14 people are accessing the channel and communicating there. We also see a rise in the use of intelligent agents—chatbots driven by large language models. There have been significant harms associated with using these chatbots, particularly for children. In some of our testing, we have seen how easy it is to produce potentially radicalising content from quite a benign prompt being put into a chatbot, so those kinds of narratives might start to justify extremist views. Young people are using these agents across a range of contexts.

I completely agree. Going back to the first question about that ecosystem, which includes the end-to-end encrypted decentralised platforms, we also find that the file-storing websites are a key component. You end up with a kind of triangle where they populate the large-scale social media platforms, either trying to stay just on the right side of the policy line, or using it as spam places to put a lot of links. Then you have these end-to-end operationally-secure ones, and they all post links to the file-hosting websites. Whether that is Islamic State propaganda, footage of the Christchurch shooting or whatever, it is an easy place where they can be relatively sure that it will evade content moderation. The big problem with those file-storing sites is that the barriers to entry to setting one up are so low compared to all the other online platforms. All you need is some bandwidth—quite a lot of bandwidth—to put it up, but compared to setting up a social media platform, particularly an end-to-end encrypted one, it is much lower. I have talked about it already so I will not go on, but it has been demonstrated that the artificial intelligence that is built into the platforms’ recommendation algorithms can promote and amplify progressively more extreme content.

In a landscape where serialised propaganda as a component of extremism and terrorism can be digested by young people especially, it is particularly concerning that it is much more prominent now than it might have been two decades ago. Thank you; that is really interesting.

How big are these groups on the end-to-end encryption?

It is not so much how large they are individually, but how many there are. They vary depending on the level of content being shared and if it is very extreme. It is not just ideological content or content around violence; child sexual abuse material is also shared on these sites. The more extreme the material, potentially the smaller the groups are, but there are many of them. Some are more present to share other types of content over another. They have different purposes. They are broadly known as a network where there are thousands of people involved, but because they are decentralised, there is not one group with one number associated with it.

The people in these groups tend to be very tech-savvy. They will have plans and many back-up plans. With the Islamic State on what was Twitter, we are almost going back 15 years to the games of whack-A-mole we were playing. We might talk about the comm networks later, but Telegram will occasionally just swipe some out. But these are smart people; they know the next place to go and these things are back up and running within hours.

You have mentioned AI and chatbots and I suppose there are opportunities in terms of how we tackle this problem with AI. In terms of the balance between whether it is proving more useful to people trying to tackle it or to those trying to spread extremist content, I am guessing from what you have said that perhaps the balance is not in our favour at present.

It is difficult to say because we do not have the data to answer that question, but the issue is not being tackled with existing measures.

With the ease and ability to create, spread and reinforce material, is AI having a significant impact on radicalisation?

Yes, and to create the material.

There are some ways in which AI is being used by platforms. The primary way in which AI has been used in the fight against extremism is by identifying and removing terrorist content before it is even uploaded.

Illegal content.

If you had someone from the tech companies here, they would say, “We are all part of a big conglomeration. We all hash share, so if someone tries to upload content, we get a big amount of it down before it is uploaded.” However, none of the sorts of platforms that we have been talking about over the last few minutes are in the Global Internet Forum to Counter Terrorism, so they are not using that kind of technology.

We have touched on this, but could you reflect a little on the link between, or any interaction between, online and offline spaces in spreading extremist narratives? Dr Whittaker, you have written about some blind spots within Prevent and the difference between how it works offline and online, but it would be interesting to hear your reflections on that.

The starting point is that we have spent a really long time debating things such as online versus offline radicalisation, and the truth is that it is a false dichotomy. To this day, there are very few people who simply stay in their bedroom and become radicalised to the point at which they want to commit violence without having any kind of ideological network or support. When it comes to terrorist attacks, the new normal is lone actors, but these individuals almost always have offline networks of support. It may have been the family they grew up in, their local community, a social setting, or something like that, and these things bounce off each other. Again, it was quite common with the Islamic State for a group of friends to go to someone's house and stream radical YouTube videos together. Is that an online or offline behaviour? The distinction does not really make sense today. When it comes to countering it, it is clear that for all Prevent's benefits, it does not really touch the online space in any way and that is the primary way in which young people are communicating with each other today.

On that point, we are data poor. We do not have large enough datasets that triangulate across online and offline factors to be able to answer that question in a valid way.

I would like to add that the reality is the distinction between online and offline is a false dichotomy not just with regard to extremism and radicalisation but with regard to everything. If you think about how your own social or professional life works, you will be WhatsApping people you saw in the corridor five seconds ago. Everything constantly works between the online and the offline spaces. The online now permeates the offline to such an extent that you cannot actually divide them. It is important to recognise that where somebody is in that situation, perhaps by themselves, accessing content that is perhaps over the edge a little in a private space, not directly interacting with anybody else whose views are extreme, certainly not part of an extremist group, these people are sometimes mistakenly pulled into the counter-radicalisation space. They might be treated as a Prevent concern when in reality someone who may have, for example, an autistic special interest is not necessarily linked to anything they are going to do in the real world. As well as paying more attention to the online space, we have to look at what people are actually doing with this content, and what their intentions and motivations are. In some cases, they really should not be looking at that, but it does not mean this person is necessarily radicalising towards violence.

I suppose the question is about differentiating between someone looking at that sort of content and the action they wish to take. That feels like a much deeper dive. From a national security perspective, I guess the question is whether you would still want to flag that person and how that would work in practice, but I accept the premise.

If I could add something: the original purpose of Prevent was to prevent people from being drawn into extremist groups and networks. It was not just for people who were maybe accessing a little content that they should not. It was actually to prevent people from being drawn into these networks which involve other human beings who are more into the ideology than they are and who will guide them towards violence. That was what Prevent was supposed to stop.

I understand what you are saying, but in practice when we have a system where even repeat referrals sometimes do not flag someone who will ultimately go out and commit serious action, the question is about balance, is it not?

It depends what we mean by new forms of extremism. Some things which have been floated as new forms of extremism are not forms of extremism and many of them are not terrorist threats. Generally, the response in counter-extremism should mirror the threat picture, and Prevent clearly has not been doing that. The Shawcross review very clearly highlighted that the majority of referrals are related to so-called mixed, unstable and unclear ideology, where essentially there is no terror threat. The second largest is the extreme right; the third largest is Islamism which is, in fact, the greatest terror threat that the country faces. There are new forms of extremism, but I have not seen them discussed in, for example, the leaked counter-extremism sprint, which was carried out. For example, there is a serious and growing problem in this country with anti-blasphemy extremism, and as I have already mentioned, there is a serious and rapidly growing problem with anti-Israel extremism. Both issues are used for radicalisation by Islamist groups and by non-Islamist groups that cannot be described as Islamist. They are not trying to establish a caliphate; they are not related to the Muslim Brotherhood or its Shia equivalents, but they are using this to rile people up and essentially to radicalise people towards violence in some cases. One of the main victims of the anti-blasphemy movement in this country has been Ahmadi Muslims, who are treated as heretics by some other Muslim groups. As far as I can make out, these forms of extremism are not being addressed by Prevent at all. My worry is that Prevent will end up wasting time looking at things like the incel movement, which is problematic and needs to be looked at but it is not a form of extremism, and there is no terror threat in the UK.

For the most part, yes. I would probably argue that the incel movement is an extreme movement, but that is a minor disagreement. I am going to take it a slightly different way and talk about who is under the Prevent duty and how well they are trained. In answer to your question, Prevent is not very adept at identifying new forms of extremism because the current level of training for the people who are under the Prevent duty is seriously lacking. I tried to look for it. I am not sure exactly how many millions of people are under the Prevent duty, but for almost everyone these days it usually includes an hour's online learning and then a quiz the first week that you start a role, and it is very rarely revisited after that. When I did mine, it was mostly focused on the Islamic State. I am not sure a lot of the people under the Prevent duty would have any idea about comm networks or incels, or even things that relate to Israel and Gaza post-7 October. It is problematic in terms of how you allocate those kinds of resources to make sure that people are adequately trained. At this point, I am not sure I have much faith in the general competence of the referees. I do not mean that in a rude way, but simply that if they do not have the time to dedicate to a two-hour session once a year or something like that, I do not think they are going to be particularly au fait with all the new stuff. Problematically from that, public events can really sway what it is people are referring to. For example, a popular Netflix drama may massively change the way in which teachers, social workers, and health workers are referring people. I have no problem with “Adolescence”, but it is not training for Prevent, and it is potentially taking people in other directions that are not backed by evidence.

I agree with Dr Whittaker's points. Just to speak more about these comms networks: they pose an issue for Prevent because of the definitions around terrorism and extremism. However, referrals are made to Prevent because people are getting involved in these networks, and that is because these networks are involved in sharing terrorism-related material. Some people trigger a Prevent referral because they have shared or accessed some material, but the role of ideology in these networks is really unclear. They are sharing neo-Nazi content, but ideology does not appear to be the driver. People in these groups are not trying to change the world order in some way; it is more of an aesthetic. That is something where more investigation is needed. It is a tri-threat area. There is terrorism-related offending, cyber offending, which could be attacks on national infrastructure, and child sexual abuse and exploitation in these networks. The problem is that different agencies on the front line may see one, two or three of those offences and not know whether to refer or not, and it is a grey area. Members of these groups may be relevant to Prevent, but not in the same way as we might have seen other groups being relevant. Addressing the ideological drivers might not be appropriate to disengage them from these networks either.

Yes.

This is a really controversial debate because at no point do we want to stigmatise people who suffer from mental ill health or who are neurodivergent. That is the first point I want to make here. There is no data that provides a causal connection between neurodiversity and terrorism. It is not a risk factor for the general population, and there is research that I can forward to the Committee on that. However, there are some hypotheses around how the specific facets of certain neurodevelopmental conditions might relate to engaging with extremist content and so on, but like you said, more research needs to be done there and more data on those issues is needed. Mental health and neurodiversity are two umbrella terms. Mental ill health captures a range of diagnoses. There are multiple reasons why people with different types of diagnoses might have had, for example, adverse life experiences and trauma, as well as mental ill health. We cannot disentangle those as reasons why they then become involved in a group because of a sociopolitical grievance. Understanding the intersections between those factors that give rise to susceptibility is really important. There are some writings on different neurodevelopmental conditions, but no data at the moment that gives that causal link.

There is a very small base of empirical research. We use the word neurodiversity, but when we are talking about links between extremism and neurodiversity, we really mean autism. There is very little research on ADHD or any other kind of neurodiversity. The research that exists is thinking about autism. There are about three empirical studies which have generally found something of an overrepresentation within terrorist or violent extremist samples, but that does not mean it is causal. That is really important to underscore because roughly 1% of the population has autism, and it is a fraction of those people who actually commit violence. Those kinds of hypotheses are usually that it is autism in connection with something else, some other kind of life experience, things that relate to impulse control, special interests, things that might be online activity and might put people in a certain space which might end up in their radicalisation.

I very much agree with what has just been said. I want to add something though, which is that there is a lack of rigorous empirical research, not just on the idea of a link between neurodiversity and radicalisation but with regard to everything in the extremist space. There is a terrible lack of research. I did a survey of the national research environment for the study of extremism for the Commission for Countering Extremism which came out a couple of years ago. I found that there are immense barriers to carrying out research on, for example, convicted terrorists. In fact, I am aware of only one person who has been allowed to do this work by the Home Office. They carried out a particular project that they were commissioned to do, and that was very important work. There are lots of people who want to do research by interviewing convicted terrorists to try to understand the pathways that led them into extremism and violence, but it is immensely difficult to do. The Prison Service itself does not co-operate. This research has to be done. There are people doing that research in France, for example, but there is nobody really doing it here. We need that work.

I alluded to this point earlier. In my mind, Prevent would be best suited as a very narrow counter-terrorism point. It would probably end up as a better outcome for all involved if there was this broad entry point with lots of different offshoots: possibly counter-terrorism, violence fixation, and a funnel to neurodiversity or mental health support. If it was deemed relevant to have a counter-extremism aspect, or you might call it integration—whatever you wanted to call it but something that is not counter-terrorism—that had some kind of intervention, then the training would say that the referees at the front end, the teachers, the academics and the social workers, would merely need to identify some kind of problem and then refer it to their safeguarding lead, rather than the situation that we have now in which Prevent is the only game in town if you see some kind of problematic behaviour. That is a slight exaggeration, but for the most part, it is the easiest access in, again, that hypothetical example where a teacher has problematic student behaviour in class and they say something that can be construed as extremist in some way. I do not blame the teachers because that child needs support and they need to stop disrupting that classroom, but it is really problematic for all involved for that child to go through a counter-terrorism scheme.

I wholeheartedly agree with what has just been said. I want to add that there is a problem with Prevent training. Obviously, I am going to say that there should be much more focus on antisemitism as a component of extremism within Prevent training materials, but that is a specific problem. The general problem is that Prevent training is generally provided through very poor-quality online materials. The nicest thing I can say about the online Prevent training I received is that it was probably no worse than the online fire safety training I received. These short online courses with their little quizzes to test whether you have done your learning are not adequate for anything, and certainly not for helping professionals understand who is and who is not at risk of radicalisation towards terrorism.

Where there are emerging types of threats from new groups that have different types of behaviours associated with terrorism, alerts should be put out to frontline workers so they can recognise those signs.

There has not been a lot of work done in the online space for those permissive environments and people's networks online. If there was some kind of online youth club equivalent, that would be helpful for disentangling those networks.

I have little to add except to say that for a younger child to go through a counter-terrorism programme is potentially life-changing and very unlikely to be in a good way. Almost certainly, given what we know about the Prevent demographics, there needs to be much more emphasis on the care and support elements rather than necessarily thinking deeply about the ideological challenge at that younger age—not that it should be removed entirely—and potentially it having a different title. Something that is not explicitly counter-terrorism would probably go a long way, particularly as we are talking about children.

I second that. There are far too many people being referred to counter-extremism interventions who are not in any sense terrorists or moving towards any sort of terrorist involvement, especially children.

Possibly, yes. At the criminology department, a lot of my colleagues would say that diversions, as they were once called, have a lot of promise. There are a lot of concerns around funding in the criminal justice system and how that would be rolled out. If there is a way of doing that that ensures rigour by having good funding, then it could absolutely have some legs.

The fact is that this work has not been done, as far as I can make out. There is no clear idea of what success means. There has been no research that I am aware of looking at the effectiveness of interventions in this country. Dr Whittaker, maybe you are aware of it?

Not in this country, no. I was about to make a more global point.

No, I am making a point about this country. The work has not been done. That is one of the many aspects of extremism in the UK which is criminally under-researched.

Most people in this sphere call themselves a CVE—countering violent extremism—practitioner. There are a lot of people who work professionally there and they will tell you that today it is basically impossible to get a big grant to do one of these counter-extremism or countering violent extremism projects without a very robust evaluation plan at the point of writing the grant bid. There is certainly nothing public to suggest that is happening on any level with Prevent. There is this fundamental issue that anything involving violent extremism needs to be evaluating the counter-factual. How do you know that someone would have become radicalised, but they did not? There is a whole host of different types of evaluation that you can do, which can be anything such as attitudinal surveys as you are doing it or, if children are involved, Prevent will involve a parent, so ongoing discussion about how children's behaviours and attitudes are changing. Then there is what you might call a process evaluation to make sure that the actual process of the intervention is being done in a rigorous way. It is entirely possible that there are elements of that happening in Prevent, but those are not things that are released to the public at any point, so as it stands it is a bit of a black box looking in.

I agree. Obviously, there are a few social desirability concerns when we are asking people what their attitudes are. Trying to establish whether somebody's attitude has really changed is almost impossible, but I would say that looking at somebody's social situation, the networks and groups they are involved with over time and whether that looks apparently non-extreme and healthy is a good metric to go with.

You made the point that perhaps this research is being done internally by Prevent and not released. If research is done internally and not released for critique by the scientific community, it might just as well not have been done at all. Science has to be public; it has to be open.

I am going to bring in Chris Murray for one very quick question and then we will wrap up.

Thank you. I want to go back to Prevent for a second. There is always some discussion about whether Prevent needs to be reformed or whether it needs to be replaced in some way. I want to test two hypothetical arguments on you. One value of Prevent is that it is now quite embedded. Teachers know about it and those under the duty are aware, maybe a bit too vaguely. Do you worry that if we had some big reform or replacement, we would lose that one advantage? Secondly, do you have concerns about the national debate we could end up having if we decided to completely reopen the idea of Prevent and how we would structure that debate? Would it lead anywhere productive? That is not meant to be a leading question although I realise it sounded like it.

Do not scrap Prevent. Generally speaking, getting rid of an institution and trying to start again from scratch never leads to anything good. It is much better to reform things than to destroy them and try to make something new.

As I suggested, perhaps there could be a model with multiple outputs, of which Prevent is one. You could think of a different name for the overall safeguarding approach, or just call it safeguarding if you like. Potentially, that would go a long way. There is a lot of debate about the claims of communities feeling over-securitised. One of the ways to potentially address that is to have fewer people from those communities inappropriately going through because typically only about 10% of Prevent referrals make it to a Channel panel; the rest are considered not appropriate. That would have an effect on the trust in communities if a lot of people were being identified by teachers and social workers and then deemed not to be appropriate. Something like Prevent still existing would keep the benefit of teachers and social workers knowing the sorts of things to look for, but it not being the only game in town.

I have nothing to add.

I will bring this panel to a conclusion now. Thank you for your time and your willingness to answer in very open terms. It has been fascinating. I will suspend the sitting now for a couple of minutes while we change over the panel. Witnesses: Robin Simcox and Lord Anderson of Ipswich.

Welcome to our witnesses for the second panel on our session today on tackling new forms of extremism and radicalisation. Could I ask the panellists to introduce themselves, perhaps starting with you, Mr Simcox?

Thank you, it is great to be here. I am the Commissioner for Countering Extremism, which means I am the Home Office's independent adviser on extremism. I have been in the role since March 2021.

I am David Anderson. I am a barrister and a Cross-Bench peer. I am also the interim Independent Prevent Commissioner. The one thing I would like to say before we start is that, as chance would have it, my report on lessons for Prevent is going to be published tomorrow morning. That might perhaps make it a bit frustrating for you and certainly for me because I cannot tell you what my conclusions are, although I do not see why we should not discuss some of the issues that are well known to arise in relation to Prevent.

Thank you very much. We will start the questions with Chris Murray.

Thanks to you both for coming. Can I ask a very direct question at the beginning, which is: what is the difference between your two roles, and how do you work together or work separately?

They should be complementary. It is true that since I became commissioner, the CCE has taken on more Prevent-facing work than under my predecessor. After the independent review of Prevent, I was involved in some way, shape or form with having a level of oversight over half the recommendations made. We also run the Standards and Compliance Unit, which is a body that has independent oversight over bits of Prevent delivery that we imagine will soon be integrating into the Independent Prevent Commissioner. It is right to say though that the CCE does not have to be involved closely in Prevent. The extremism field is not defined by its relationship to Prevent. There is an awful lot more that you can discuss around this that you can do work on that does not relate to Prevent, and I am sure we will get into some of it today. The commissioner can do weather setting about the state of the extremism challenge they face. They can probably get more involved than Prevent may wish to in upstream social cohesion, community cohesion-type issues. So there is the potential for crossover, but in practice the two roles can co-exist very peacefully, and so far have certainly proven to do so.

I see my role as very practical. It is really similar to the role of Independent Reviewer of Terrorism Legislation that I used to do some years ago. I am there to review Prevent and bring as much as possible out into the open. I very much agree with your last witnesses about that. I talk to as many people as I can and try to respond to their concerns wherever they are coming from. I pull up Prevent where it is going wrong, and just as important, give approval where it is merited. I see myself as a servant of Parliament. Sessions like this are really important because ultimately scrutiny belongs to Parliament, and people like me are there to help you do that as effectively as you can, in this capacity at least.

Mr Simcox, how much of your work is looking at the challenge of new forms of extremism? You said that you focus more on Prevent now than in previous incarnations; do you think new forms of extremism are driving that?

I do not think they are necessarily driving it. To be effective in the role you have to be constantly horizon scanning and seeing what challenges are coming around the bend. I should first state that just because there are new forms of extremism, it does not necessarily make them more serious or more pressing. We have had a pretty well-established threat from the extreme right-wing and Islamism in this country, and as commissioner it is important I do not take my eye off the ball on those because there are new challenges and threats ahead. From my point of view, the key thing is to be flexible in the work, have an open mind, and be intellectually curious about the challenges ahead. In the four years that I have been in the role, we have had Gaza and the fallout; the incident of the terrible murders in Southport and of course the disorder there in the wake of the murders; disorder in Leicester and debates about, for example, the role of Hindu nationalism in informing some of that disorder; blasphemy-related flare-ups; and electoral intimidation of MPs. There is so much that you need to be aware and cognisant of. It is really only by being flexible and being willing to pivot to emerging threats where you need to that you can in any way be successful in the role.

I am keen to explore the trends in extremism, and maybe Mr Simcox, I will ask you to start off. How have you managed and approached the challenge of the different emerging forms and trends of extremism in your work, and what have you learned from your involvement within that process?

I have learned that you need to be on top of the latest research, trends and data, and obviously also actually get out and speak to people to ask questions directly. You just heard Dr Joe Whittaker on the previous panel. At the CCE, we commissioned Dr Whittaker, along with two other academics, to do a study which was the largest study of incel opinion so far, at the time. This was one of those topics where increasing amounts of people were talking to me about it, so it was clearly of interest to officials, parliamentarians and Westminster generally, and I sensed that the opinion and conversations were informed more by a hunch as opposed to any proper data, or by what people were saying on social media rather than actually asking the people involved what their views were. One of the things we wanted to do with the incel research was to get a bit of demographic data: whereabouts do they reside? What age are they? What political inclination do they have? Is there a mental health component there? There was a load of assumptions around these things, but I did not feel it was really being backed up by any especially informed data, and that is absolutely vital. The other piece of research I would bring to your attention on which there has been a long-running question in my world—I have been doing counter-extremism and counter-terrorism work for getting on for two decades—is about the role of ideology and the extent to which it drives acts of terrorism and extremism. We put out a piece of work by an academic, Dr Donald Holbrook, who has access to reams and reams of mindset material, as he dubs it, of convicted terrorists. He basically has access to what they are reading and accessing. From that we were able to pull out just how much terrorists immerse themselves in ideology, discuss it, debate it and sometimes create it, and we were pretty comprehensively able to answer this question about the extent to which ideology matters to terrorists: it absolutely does. So I hope that answers your question in some way.

To expand on that, and I will put this question to both you, how much, if at all, do you rely on additional data or additional themes that exist to address the growing trend of different emerging extremism?

One of the things that I have been keen to champion with Government is getting that hyper-local knowledge about what is happening at ground level. That is not always easy, and I understand the barriers and challenges that exist around it. For example, a frustration I had after the disorder in Leicester a couple of years back when I was speaking to people on the ground was that in some of the conversations afterwards trying to unpack this, I heard “Oh yes, we saw this comment months ago.” So I said, “Well, that’s interesting because we were not seeing that.” I do not think Government were seeing that being fed in. These local tensions were not really translating to this. I certainly did not have a conversation around the extent to which Hindu nationalism and local Muslim residents, with some Islamist group involvement, were driving tensions between each other and had the potential to lead to acts of violence. That was not the discussion we had prior to the summer of 2022. It has always been my instinct that you can understand a lot more about how extremism works in this country if you get close to the communities involved and get an understanding at grassroots level, so again that is a pretty significant factor.

Your question is really important, so can I tell you how I see those themes that you refer to? Some 15 or 20 years ago, terrorism was pretty easily definable. It was what dissident Republicans were still getting up to in Northern Ireland and organised plots directed or inspired by the likes of al-Qaeda, as we saw on 7/7 in London. Since then, the picture has been blurred by two developments in particular. The first is the growth of lone-actor terrorism, often using something readily available such as a vehicle or a knife—something readily available. Secondly, there is the proliferation of ideologies initially from the far-right, as we saw with the murder of Jo Cox in 2016 and some much larger-scale atrocities in Norway and New Zealand. There are a mish-mash ideologies that you have been hearing about: fragments of this and that packaged as video clips or slogans, conspiracy theories, school massacre ideation, extremist content intersecting in unpredictable ways and merging with personal grievance, anger, and sometimes a kind of rigid mindset that can be associated with neurodiversity, as was said in the last session. As a result, there are many cases now in which terrorism is no longer very different from other forms of extreme violence. The days when you could put it in a box on its own are, to some extent, in the past. It can be difficult to know whether an attack had a terrorist purpose or not, particularly in advance and sometimes even by the end of a trial. What we are seeing is that attacks that certainly look a bit like terrorism are being prosecuted on a different basis. The Southport killer—one of the cases that I was tasked to look at in the review that will be published tomorrow—had been manufacturing ricin and had downloaded the al-Qaeda training manual which he found on the US Air Force website, but the judge deliberately said, “I cannot sentence you as a terrorist because there is no evidence that you had a terrorist ideology.” So when it is said that there is no evidence of people with mixed, unclear, or unstable ideologies committing terrorist offences that might be strictly true, but they still commit offences that look very like terrorist offences. It is not just him, you could think of lots of other examples: Damon Smith, the tube bomber; Danyal Hussein, who murdered two sisters in Wembley; and Jake Davison, who killed five people in Plymouth including his own mother and had been on an incel website that morning. None of these cases is simple, none was characterised by the prosecution as terrorism, but they all come out of this developing phenomenon. They make it very difficult to draw a bright line these days between terrorism on the one hand and other forms of extreme violence on the other.

What trends are you seeing at the moment in how online spaces are used to spread and reinforce new forms of extremism?

I can start if you like, although I suspect Robin is more expert than me. Let me just give you a couple of quotes. One is from the Prevent duty guidance of 2023, produced by the Home Office. It now says, “The internet has become the ‘preferred’ avenue for those searching for terrorist propaganda or contacts”, and it records the decline of what it calls the previous hybrid pathway, including both online and offline influences. So there is a strong picture there of a large number of things going online. Then the director general of MI5 gives an annual lecture about the threat—the last one was in October of last year—and he said, “It’s hard to overstate the centrality of the online world in enabling today's threats.” That entirely translates into my experience of Prevent, including cases which do and do not have a specific terrorist ideology. Let us not forget that the published figures show the average age of a person referred to Prevent is now 16, and 40% of them are between 11 and 15, so you are dealing here with digital natives.

How effective do both of you think Prevent is at identifying and addressing online extremist behaviour overall?

I will respond to your first question first because it might help answer the second question. It seems to me very obvious and undeniable that the internet is an accelerant. It has been an unbelievable boon if you are a terrorist group that wants to disseminate your propaganda and do so fast. The scale of it is obviously unrecognisable from how this sort of content was passed around in previous decades. Of course, sad to say, terrorist groups still have a presence on some quite mainstream social media platforms. If you look at some of the content online, especially some of the most incendiary rhetoric, the challenge for intelligence services and others is being able to distinguish between people mouthing off online and an actual actionable threat that we need to do something about. I would suggest that I was going to be a slight outrider, but the previous panellists had a similar view that we have to be mindful about this idea that the internet is the only game in town when it comes to radicalising people. The evidence for that is not as firm as is sometimes assumed. I do not know if you have ever heard the phrase, “I have never met a happy bully.” I do not know if I have necessarily ever met a happy extremist. I do not buy that a very well-balanced, stable, high-functioning person who is thriving in society then sees some content online and is an extremist. It seems just as plausible to me, though it would need to be equally tested, that the internet and content online can perhaps give greater coherence to views that are already being weighed up, contemplated by the individual involved; such content perhaps brings greater coherence, greater structure. The idea of radicalisation being a solely online phenomenon assumes that you are dealing with blank slates, and you are not. You are dealing with people with all sorts of things going on in their lives who have their own agency. That is why I have never been wild about the phrase, “People being drawn into terrorism.” People also choose this stuff. People have agency, and sometimes the focus on the online component only takes agency away from the conversation.

Could I answer your second question? Again, if I may say so, it was a great question. Frustratingly, I cannot tell you how well Prevent deals with the online threat; you will have to look for my report tomorrow. But may I deal with your question by reciting something that is a matter of public record? Axel Rudakubana, the Southport killer, was referred to Prevent for the third time because a teacher in a bricks and mortar classroom physically noticed him doing a search for London Bridge on a school computer. She saw the tab and she went over and spoke to him, and as a consequence of that conversation, a referral to Prevent was made. My rhetorical questions are, how many other searches had that boy been making in the privacy of his own home on his own devices? How many people had he met online? How many contacts had he made? How many forums had he contributed to? How well-equipped was the Prevent duty to know about that and to deal with it? I am not suggesting there are easy answers, but that incident really highlights to me just what a pertinent question that is.

There have been wide-ranging criticisms of Prevent overall, and I definitely take your point about people being more susceptible if there are things going on outside what they are seeing on their phone or computer. The question is more geared towards whether Prevent is equipped enough or is doing well enough at tackling what they may see online that might radicalise them and the issues they are facing at home or school, for example, that may lead them to a place where they are more susceptible to these things.

There are a couple of ways into this. Some of the content people are looking at might be illegal, in which case there is a content takedown component to that. Generally, that seems to be one of those areas where the state, in conjunction with the social media companies, has actually been pretty effective in the last 10 years or so, let us say, since people really became aware of this when ISIS were in their pomp and so much of this content was on social media. I suppose one of the challenges is that a lot of the content which may be looked at is not necessarily illegal. Rudakubana was interested in the Rwandan genocide; he was really obsessed. Lord Anderson just talked about him googling London Bridge. There are all sorts of things you can access there and ways in which you might be able to immerse yourselves in violent content of some form that would not even be approaching being illegal. You can read all sorts about the Rwandan genocide that you would not want to be taken off the internet, for example, or you would not necessarily want to make it harder for people to access. So that is where Prevent, law enforcement, and intelligence services have a really tricky job on this. I am not even sure if it would be fair to call it a fine art because judging who, in accessing that content, could be the next terrorism or violence fixation risk is difficult; it is even hard to know exactly what indicators you would begin to look at.

I will go to my next question about the link between the online and offline space, and this may feed into some of what you have pointed to in terms of incel culture. Do you think you can identify certain individuals who may be more susceptible to certain things online? Have you seen certain individuals being targeted because of what they might be looking up in particular and then being taken down a rabbit hole which leads to more extreme views?

I share the views that were echoed in the previous session about this online and offline distinction becoming much more blurred. It is increasingly hard to sustain, and especially as AI really kicks into gear, it will become even harder to sustain the idea that they are two separate entities. When I first started accessing the internet, I had to kick mum and dad off the phone and use a dial-up modem. That was quite a clear online/offline distinction. It is obviously a completely unrecognisable world now, and I have to be honest and say that sometimes we in the policy world, Whitehall, and SW1 generally, probably think about this distinction more than young people do. I sense it that in 10 years the conversation will be about how it has broken down entirely. I will come back to my previous point about the need to focus on the local factors alongside the online. The human dimension cannot be overlooked, nor should the idea be promoted that by looking at someone's content online, you might be able to identify who would be an appropriate Prevent referral. That seems to me untested, although maybe you can. I would not want to say it is impossible because I do not know if it entirely is—although I do not know how you would do it—but that does not mean there are not much smarter people than me at the Home Office figuring out a way of doing so. Coming back to some of the points discussed earlier about how terrorists tend to form in geographic clusters, we saw this with the ISIS-related foreign fighter travel. People in pockets of the country and throughout western Europe tended to go together. If the internet was so effective in radicalisation that it did not really matter where you lived because you were just as vulnerable to being radicalised as anybody else, then you would not see that geographical clustering; it would be far more diffuse. So again, the internet is a vital, key component and a key way of understanding the threat, but it is not the only game in town.

I have spoken to community support organisations working on the frontline, reaching out to troubled students in schools and giving them mentoring. The phrase they kept coming up with was “digital pandemic”. A number of people who thought about this during our covid pandemic thought that there was perhaps a similarity, because we needed an all-spectrum response for covid, and it is a bit the same with this pandemic. You have to do what you can with Prevent to actually treat people who have the virus—try to take them off the path, cure them and put them right—but that is never going to be enough on its own without doing two other things. One is building a resilient population, and you could say that is like vaccination when you are talking about covid. When you are talking about the digital world, this means things like community cohesion, tolerance, critical thinking and digital literacy and trying to build up these skills. Then you also need a safer online environment, which is closing the web markets and stopping the spread of the virus. In digital terms, it is about regulating these platforms, getting them to change their algorithms, take stuff down and mend their behaviour. All these things are difficult, but my point is you need all those three aspects: a resilient population, a safer environment, and personalised attention for the people who are affected.

Thank you very much. You might be right, because if our search history as members of the Home Affairs Committee was to be looked at, we ourselves might be flagged. In that vein, do you think there might be way to future-proof against rapidly changing technologies when it comes to the way in which platforms are being encrypted? People are spreading a lot of hateful and extremist messages, and the tools and tactics that we have are often outsmarted by the next week. Could you comment on ways in which we could manage that?

I will be honest; I am afraid I do not feel qualified to answer that question. Technology is moving so quickly, and we are now doing things with AI that I was not even aware were on the horizon six months ago. Everything is in flux with this at the moment, and I would not be the best person to answer your question.

There is an organisation within the Home Office called RICU which does open-source research and presents really useful briefings. Most of them are likely classified, and really the more people who can see briefings like that the better because, as you say, things change very quickly. Robin is right that they do change in regions: there are clusters and there are regional variations, which is important.

I have some further questions, but I might come back to them.

Okay. We are about to vote. I am going to suspend while we vote. We will come back in 15 minutes. Sitting suspended for a Division in the House. On resuming--

We will resume our panel now, and I will turn to Chris Murray.

You have both alluded to this a little, but one of the developing trends is of extremist people with no really fixed ideology or with a mixed or unclear ideology. It is something the Home Secretary has referred to in every statement she has given on extremism since taking office. Especially given that the average age of a Prevent referral is 16, do you think Prevent is set up for or appropriate to deal with that kind of mixed or unclear ideology?

Guidance went out to Prevent practitioners from counter terrorism policing and the Home Office as long ago as June 2019, pointing out that there were people with an obsessive interest, as they put it, in public massacres and multiple concurrent and even contradictory extremist ideologies or causes, and they said that those sorts of people were eligible to be referred into Prevent. An awful lot of referrals were made for people in that category at the time, and it has been argued that one of them was Axel Rudakubana, the Southport murderer. The issue for Prevent is whether those people are best dealt with in Prevent or somewhere else, and if so, where? I do not think anyone believes that these people are not potentially dangerous enough to need dealing with somewhere. There are various options on that, and some have already been addressed in the last panel. I would love to tell you where I come out, but I am afraid that will have to wait until tomorrow.

We are looking forward to tomorrow; we will get it then.

This debate has been going on for some years. The general drift—and it has been a drift as opposed to a decision that was debated too much, at least certainly from a public perspective—is that Prevent is going to take on VFIs, and if that is not the perfect place for them, it is the most suitable place at the moment because of a lack of alternatives. I would say a couple of things on this. First, that may be the best outcome. You may all discuss and decide this is the best outcome and it will be debated among officials; obviously, Lord Anderson is doing diligent work on it as we speak. If that is where we come down, Prevent is not really a counter-terrorism programme anymore; it is a violence reduction programme with a bit of counter-terrorism in it. The trend at the moment is for Prevent to take on fewer ideological cases, and more referrals around what used to be known as mixed, unstable, and unclear—now different categorisations, but broadly what we could understand as including a large amount of those with violence fixation ideas, or perhaps violence fixation interests. That seems to me to be a pretty fundamental shift in what Prevent is, which is part of the counter-terrorism strategy, one of the four Ps, and a whole load of consequences stem from that. If that is the direction of travel, I think that Prevent had better brace itself for an awful lot of referrals, because the terrorist or potential terrorist cohort in this country is relatively small, thankfully. Those with a broader interest in violence will be a bigger issue, and the system is going to have to be able to take the strain of this potentially large amount of referrals we will see. We sometimes think about Prevent's reputation and the battering it has had because of the fact that Rudakubana was referred into Prevent multiple times, and of course he is not the only one that has been on Prevent's radar that has carried out an act of violence in recent years. I thought Prevent was in the needle-finding business, but the risk is that it gets into the hay production business, because you are going to generate an awful lot of referrals, and probably of a lot of people who do not really need to be in Prevent. If there are more and more traces on the Prevent system, and it is harder and harder to discern who is going to carry out an act of violence, that has potentially damaging consequences for Prevent's reputation, because there are going to be more and more cases where there was a referral but it was not followed up. Of course, that is going to happen all the time because knowing from the referral process who is going to carry out an act of violence is not an exact art. We do not have any ability to look into the future. As I say, at the end of the debate it may be that this is the direction of travel for Prevent, and it may be the optimum outcome of all the options available, but we should fundamentally acknowledge that Prevent was meant to be a counter-terrorism programme and, if that is the outcome, it will not really be a counter-terrorism programme .

Okay, that is really interesting; so you are saying there has been this big shift, that mixed and unclear ideologies are part of it, and that there are negative consequences of that. It sounds like you are saying you disapprove of that shift?

No, not necessarily. When you see something as horrendous as what happened in Southport—and as Lord Anderson pointed out, Southport is not the only example where someone with a violence fixation has carried out an attack—then very clearly there is a need to try to prevent that happening again, and the consensus at the moment seems to be that Prevent is the best way for that to be addressed, or it is the most mature system with the ability to deal with it. Hopefully Lord Anderson will come in with definitive answers on this that will help guide the debate. I would just point out that there has not been too much debate prior to Lord Anderson’s report, and I personally think the independent review by Sir William Shawcross a couple of years ago about Prevent basically becoming this violence fixation programme and not really doing that much pure CT went under the radar a bit.

One of the issues is that very few Prevent referrals lead to a Channel panel, as we have touched on. How concerning is that? Could it actually be a sign that Prevent is working because lots of people are referring in and they are being addressed properly, or is it a sign of dysfunction?

In a sense it is one answer to the very fair point that Robin has just made, that you might end up with an awful lot of referrals. The reality is that fewer than 10% of those referrals end up with people in Channel panels. Pretty consistently, certainly over the 10 years until 2024, which are the latest figures that are published, you are looking at somewhere around 7,000 referrals a year and around 500 people going into Channel; in total I think 5,000 people have been helped by Channel panels since 2015. Those figures are in a sense controllable in that way. The fact that only 10% go into Channel does not mean that the other referrals are a waste of time. A lot of them, of course, go into other services, whether it is drugs and alcohol, rehabilitation, mental health, special educational needs or children and social services, but some go up to Pursue, which is the tough end of terrorism, where Prevent is not enough for them because they are already suspected of committing criminal offences and they need the attention of counter-terrorism police. The only other thing I would say about this is that if you are prepared to look at Prevent in a fair way as a safeguarding programme for individuals and the rest of us, among other things it is actually quite usual in safeguarding programmes for there to be far more referrals than actual disposals. If you look at referrals to children's social care in 2023-24, there were about 620,000—a totally different scale from Prevent—and yet you get far smaller numbers than that who end up on child protection plans. I do not hear people saying it is a scandal that all those people are being referred. There is an element of better safe than sorry, which is not entirely out of place, particularly when you are dealing with such a serious risk.

I definitely understand the reasoning behind the question, but it is not one of those things I am especially concerned about, because I do not think a higher number of people being put into Channel would necessarily be a sign of a more sophisticated and mature system. It seems to me that Prevent referrals can reflect the uncertainties, biases, inadequate training or whatever it may be of those making the referrals. In the post-Southport world, and with, I would suggest, a little bit of influence from “Adolescence”, the Channel chairs and statutory partners are going to have a lower appetite for risk. It is a hypothesis; we will have to see whether I am right or not, but I think you will start seeing more people being referred into Channel because people are going to be really nervous about not having another Rudakubana. That is completely natural and understandable, and it may mean that the Prevent referral to Channel adoption numbers start going up from 10%, but I am not sure that necessarily means you would be dealing with a lot more people who actually pose a risk, it might just be a reflection of the zeitgeist of the times and the caution people feel about not getting this wrong, as was the case with Rudakubana.

I want to ask about evaluating the effectiveness of Prevent, because obviously one of the huge challenges in a preventative system is that you cannot measure what does not happen; you only see when it goes badly wrong. What are your views on that? You have already mentioned that Prevent gets a kicking quite a lot of the time. How do we evaluate it and properly measure whether it is being effective or not?

I agree that is super-important and difficult to do, because every time Prevent fails it is literally front page news around the world, and every time it succeeds no one gets to know about it because the person has gone off to live a blameless life, or at least a life that has not engaged the counter-terrorism machine. There are other difficulties in doing a really disciplined academic study of whether Prevent and Channel work. To do it properly you would want a control group, but how can you ethically have a control group of people who you think desperately need help because they might become terrorists, but not give it to them so we can do our study? You obviously cannot do it. It would also need to be a longitudinal study that would stretch over several years. I was asked to look at two cases, one of which was the killer of Sir David Amess, Ali Harbi Ali. In both cases it was more than three years between their engagement with Prevent and their attacks, so the fact that someone is all right for a couple of years is not necessarily the end of the story. So there are difficulties in doing it. There is more I would like to be able to tell you about attempts to evaluate the effectiveness of Prevent. I can assure you people are thinking very carefully about that, but again, that one is going to have to wait for tomorrow.

You have definitely whetted our appetite to read your report.

That was the idea.

Mr Simcox, do you have a view on that?

I do not have anything to add especially. It has obviously been a long-running question about how to do this and how to prove a negative. It is something that Sir William grappled with in his review and something that Home Office officials and counter-terrorism police with whom I have been speaking to are aware of the problem. The answers are as elusive as Lord Anderson has laid out.

Could I add one additional plug? If any of you are in any doubt that Prevent can work, I made a radio programme about seven or eight years ago called “Understanding Prevent” and it is still on BBC Sounds. It starts with an interview with a young ex-Nazi who explains how he was turned around by the mentor who came to play pool with him once a week in the hostel where he was living, and then comes an interview with the mentor, Nick Daines. I defy anyone to listen to those interviews and not conclude that Prevent can work when you get the right intervention provider in the right place delivering the right message. The challenge I and the whole Prevent machine face is trying to make sure that it works more often, and that is not easy, but it is so important.

Thank you, another plug well received. A final question from me, repeating the question I put to the previous panel. For all its flaws, part of the value of Prevent is it has been in place for a long time, so teachers, social workers, and those under the duty have at least a vague idea of what it is. Do you see that as part of its value and do you agree with me on that? Secondly, if we put more resources and guidance into tackling those under the duty and improving that group’s understanding of Prevent really lead to an increase in quality of referrals, or more triaging of referrals before they come in,?

Yes, I would definitely say that you should not scrap Prevent, rebrand it, or create a new structure from scratch or anything like that. I do not see that being an especially worthwhile pursuit. Prevent gets a bit of a hard time domestically, but funnily enough, when you speak to government partners abroad, they tell you Prevent is the gold standard. We should remember that sometimes. What we have here in the UK is looked at with some envy by our friends overseas. The training was touched on a little in the previous panel as well. We do a lot of work on training, and have trained about 1,000 Prevent practitioners on in-depth, multi-day training on ideologies over the past couple of years; it has been a big focus of our work. We have also done a lot of work with the Home Office in trying to improve the online offering. The challenge is always about scalability. I sit in a lot of the sessions and the training we do is fantastic, but of course there are millions of people under the Prevent duty, and we can only work with those in the Prevent network who are doing frontline work. We obviously cannot deliver this training, nor should we be able to, to every teacher in the UK. Inevitably, there is an element of clicking through with online training. To give an example of why I think in-person training matters and why it is so important and effective, one of the big debates that has been had in this area for years is over the use of the term Islamism, and the extent to which it is helpful, offensive or turns more people away from Prevent than it brings into the room, for example. One of the things I have seen again and again with the training we have done is that having a space for people in the room to address that in the first half an hour—it always comes up—basically removes any venom that may exist around that argument. People have a chance to have their say, they hear why it is not just as simple as removing the word, and why it can be very helpful in certain circumstances. There is a richness to the conversation that can then take place that it is just not possible to get online. So as I say, we have been at the forefront of the training that has taken place on different ideologies over the past few years, and I absolutely recognise its importance.

If I may, I will just add something quick on that. It is right to say that the basic training for a teacher or a health worker might just be a half an hour online course. but I do not think they are bad. My daughters are teachers, so I have done their Prevent training courses and, for the time they have, they are pretty good. But there is more to it than that: there is the notice, check, share idea. Teachers do not just make referrals on their own on the basis of their 30-minute training; they will check and share it and discuss it with the school safeguarding lead. There are some schools where it is even better than that, where they even have a social worker who is embedded in the school for part of the week, or maybe they have a mental health nurse, and it is all discussed before the reference is made. I actually came across that in Southport with a scheme called Team Around the School, which is used in other parts of the north of England. Another possibility is what they call dual piping or dual referrals, where you refer at the same time to Prevent and to your local MASH, your multi-agency safeguarding hub. So if it turns out that in fact you were barking up the wrong tree by thinking of Prevent, there is still something that can catch the person and maybe give them what they need.

To follow on from that and some other remarks that you both made about perhaps not everything fitting into the box of Prevent, is there any evidence to suggest that, in the space where some people have lost their budgets, where some organisations that perhaps lean more towards social care and youth work have less funding, lean to Prevent funding to to fill that gap, as opposed to actually having an enormous amount of concerns that these individuals may go on to lean into extremism? Has there been any evidence of that?

I am not exactly sure about the evidence on that point. I know that Sir William Shawcross's review of Prevent found that Prevent was definitely taking up the slack for mental health services, for example. For an individual who needed assistance of some sort, it was clearly the case that a quicker route to getting there was via Prevent, as opposed to very overstretched mental health services that would not be able to deal with that person in such a speedy manner. That is obviously a highly imperfect scenario because it conflates what terrorism is and is not, and because Prevent should not be being used as an outlet for those who may have mental health concerns, although there is a very tenuous link to terrorism. It just strikes me as not being the intended purpose behind that programme, let us put it that way.

Sometimes the perception is that there is money for counter-terrorism when there is no money for anything else, and I agree. There is a perception in some quarters that if you refer someone to Prevent, maybe they will get a CAMHS appointment more quickly. I do not think that is always true by any means, but Sir William Shawcross, and indeed Robin, may well be right that that could be one of the factors that is driving Prevent referrals, and possibly even adoptions into Channel.

The final question is with Margaret Mullane.

I know what you are going to say, but I am going to ask anyway. We are just wondering what the key issues are in your end-to-end review of Prevent, Lord Anderson, and if you can tell us anything about what you would like to see happen next?

I will play a very straight bat if I may. My focus is on the two cases I was asked to look at. The killer of Sir David Amess is one, and I understand there is a debate in his honour, possibly even next week in the Commons.

The pre-recess adjournment debate is always named after Sir David because he was such a regular contributor, and if he did not mention city status for Southend at least once in every debate, then something was going wrong.

I was told that today by the Shadow Leader of the House, and I was delighted to hear it.

He is much missed.

The other case was the killer of the little girls in Southport, and of course we are nearing the anniversary of that appalling incident. What I did was look at the reviews of those cases that have already happened. I reviewed the changes that have been made in response and here is a sneak preview: there have been an awful lot of them. Then I tried to identify some areas that needed further attention; whether they are the same areas that interest the Committee I could not possibly say.

Great plugs have been given. Can I thank you both for waiting and starting slightly later, and for being patient while we voted? We appreciate it. I have found this an absolutely fascinating couple of hours, and I have learned so much. I am sure we will have more questions that we may come back to you privately on as we continue with our work, but that concludes this session.