In this chapter, I begin by exploring the framing of vulnerability in the UK’s Prevent strategy since 2011 – then move to analyse very recent developments in mental health professionals’ involvement in counterterrorism and public protection. The UK has been at the forefront of efforts to integrate mental health professionals into counterterrorism. Since 2011, radicalisation prevention responsibilities have been placed on the National Health Service (NHS) as part of the Prevent strategy. The training provided to NHS staff (to identify persons ‘at risk’ of radicalisation) has been the subject of much research, exploring how healthcare ‘safeguarding’ policies have been adapted to fit a counterterrorism mandate (Augestad Knudsen 2021; Heath-Kelly 2017; Heath-Kelly and Strausz 2019; Younis 2021; Younis and Jadhav 2020). Safeguarding, traditionally understood as the protection of the vulnerable in society from abusers, has been moulded to serve a counterterrorism goal.

These studies carefully explore how ‘vulnerability’ is presented in the UK’s Prevent programme as a transitional space between mental illness, disadvantage and potential terrorist involvement. The positioning of vulnerability discourse in preventing/countering violent extremism (P/CVE) is particularly fascinating as it does not separate groups of people into static categories of ‘vulnerable’ or ‘deviant’ (as Kate Brown notes of the UK Government’s different depicters of young women who sell sex, and young men who have received antisocial behaviour orders) (Brown 2015: 6). Rather than separating categories of ‘the vulnerable’ from the ‘antisocial’, P/CVE programmes rely on notions of radicalisation that articulate a transition from vulnerability to dangerousness. In these contemporary readings of terrorist violence, vulnerability is framed as a mutable state that can signal foreboding signs of possible violence. Vulnerability is used in radicalisation discourse to frame a window of intervention on a subject before they cross the Rubicon and commit a terrorist act (Heath-Kelly 2013). It is a ‘soft-securitisation’ or ‘pre-emptive welfare intervention’ process, whereby a person is not criminalised or exposed to the intervention of security services but their radicalisation risk is assessed by a multi-agency panel (through risk assessments such as Extreme Risk Guidelines 22+ [ERG22+] or Vulnerability Assessment Framework [VAF)]) which can provide Channel deradicalisation mentoring or support from public sector agencies to meet needs.

This creates an intervening space between standardised welfare provision (which is provided according to need, not risk of violence) and security measures implemented upon persons posing a threat to public protection (and the state). Under the banner of ‘vulnerability’, a person’s needs – and the potential risk they pose – are brought together to enact prioritised local interventions, utilising the models of both welfare provision and public protection. Indeed, recent guidance from the Home Office on the Channel programme encapsulates this sectoral and functional blurring and is entitled: ‘Channel Duty Guidance: Protecting People Vulnerable to Being Drawn into Terrorism’ (Home Office 2020).

Framing persons as vulnerable to radicalisation enabled the development of a ‘whole-of-society’ Prevent programme, involving all state agencies. Rather than trying to track the activities of radicalisers, who might operate from beyond national borders, contemporary radicalisation prevention programmes frame the behaviours and emotions of radicalisers’ potential victims as a landscape through which terrorism can be detected. As such, the profession of counterterrorism extends beyond the specialised space of police units – who track terrorists – and diffuses across healthcare, local authority public protection panels, education, social services, prisons and beyond. The potential victim becomes the sign of the possible future terrorist event, as the social welfare landscape opens to perform counterterrorism work upon the vulnerable.

By focusing on the recent creation of vulnerability support hubs and the Police, Crime, Sentencing and Courts Bill 2021 (PCSC Bill), I extend the analysis of counterterrorism and psychiatry beyond the operations of the Prevent strategy. The hubs are a parallel structure, and separate enterprise, to Prevent – analysing all those referred to Prevent but not forming part of the Prevent strategy. While Prevent incorporates all healthcare staff in an early terrorism prevention role, more specialised responsibilities are placed on mental health professionals working with the hubs and through reformulated multi-agency public protection arrangements (MAPPA) articulated in the PCSC Bill (UK Parliament 2021). So, while Prevent diffuses (limited) counterterrorism responsibilities to the entire public sector, these new initiatives reclaim specialised counterterrorism roles for specific mental health and public protection professionals.

This ‘elastic band’ dynamic – which simultaneously expands some spheres of responsibility, while curtailing more intensive counterterrorism roles to specific fields of expertise – has also affected the distribution of vulnerability and securitisation in UK public protection. I explore how the hubs and the PCSC Bill enact a re-bordering of risky and vulnerable subjects, allowing professionals to downgrade large numbers of counterterrorism concerns and reserve resources for the securitised few. Prominent within this re-bordering effort, the state – and not the individual exposed to possible radicalisation – is framed as vulnerable (reversing the safeguarding logic of Prevent).

The confused applications of vulnerability in Prevent, the vulnerability support hubs and MAPPA reform efforts all speak, I argue, to economic costs and governmental efforts to increase efficiency in terrorism prevention. Since its expansion nationwide, Prevent has produced huge numbers of irrelevant referrals, all costing money to investigate and process through the multi-agency system. In response to these escalating costs,¹ the UK Government has tasked mental health professionals with identifying those individuals most dangerous to the (vulnerable) state, on account of their disorders. Yet this ableism introduces a profound contradiction in the logics of vulnerability. Where traditional healthcare safeguarding practice would treat a disordered person as vulnerable, and in need of protection, the recently developed preventive counterterrorism structures (the hubs and MAPPA reform) instead characterise the state as vulnerable to the threatening, disordered subject.

The Prevent strategy, safeguarding and vulnerability support hubs

The redesigned UK Prevent strategy of 2011 portrayed ‘vulnerability to radicalisation’ through a safeguarding frame in the health and social care sector – whereby one’s care needs could make one susceptible to the influence of abusers (Augestad Knudsen 2021; Heath-Kelly and Strausz 2019). Long-standing safeguarding policies in the NHS (and social care) respond to situations where a disability, addiction or mental illness could make protecting a vulnerable person from abuse vital. As shown in existing research, these safeguarding provisions exist to protect persons from financial, sexual and violent abuse – where a known threat exists (Heath-Kelly and Strausz 2019). Only then can an individual’s agency (or civil rights) be overruled by a multi-agency panel to protect them.

The framing of ‘radicalisation’ as a form of ‘grooming abuse’ in Prevent adjusted the performance of vulnerability. It implicitly created a new form of abuse (‘political’ or ‘ideological’ grooming) against which vulnerable people need to be protected by the state. As stated in NHS England’s Guidance for Mental Health professionals:

Here, as elsewhere in the UK Prevent strategy, the counterterrorism policy is presented as a safeguarding duty – where a person needs to be protected from becoming a terrorist. Radicalisation-abuse, by a terrorist recruiter is the process by which their vulnerability would be exploited and transformed into terrorist threat. Taking this safeguarding approach to its logical conclusion, NHS policy documents on Prevent occasionally refer to convicted terrorist attackers – such as the Glasgow airport bombers – as ‘victims of radicalisation’ (Department of Health 2011: 6). This involves a radical inversion of traditional security logics where enemies are rarely described as victims of exploitation. While it is valid to enquire into the exploitation that may have led someone towards criminal conduct, the Prevent strategy has used the safeguarding/victimisation frame to open the entire public sector to counterterrorism work. As health, education and social services deal in preventing victimisation, the framing of terrorists as ‘having been victims’ enables the importation of counterterrorism into these sectors’ duties. Vulnerability has been the matrix through which national security duties were imported into UK public services.

Framing early referrals to a counterterrorism programme as safeguarding duties also resulted in severing the link between safeguarding and formal care needs. Previously, adult agency was protected as a rule – even when the individual was likely to make bad decisions – because liberal democratic structures respect individual autonomy. The Care Act 2014 legislated that only those adults with a formal ‘care need’, who were also assessed as being unable to protect themselves from abuse, could be intervened upon by safeguarding structures (UK Parliament 2014). Since Prevent was introduced as a safeguarding duty, local authority safeguarding has been enacted by professionals in the absence of relevant care needs (including disabilities, addiction or mental illness). Undermining the traditional centrality of care needs to safeguarding, everyone is presumed vulnerable to radicalisation under the safeguarding logic of Prevent (Heath-Kelly and Strausz 2019). Home Office guidance on Channel processes confirms, relative to thresholds in normal safeguarding, that:

Prevent radically reinterpreted vulnerability. It is no longer a formally recognised, socio-medical propensity to suffer abuse – but has become a novel and securitised framing of the entire population as at-risk of subversion by malign actors. While one’s formal care needs (disability, mental illness, addiction) remain central to ‘normal’ safeguarding against financial, sexual and violent abusers, it is only Prevent where the centrality of formal care needs has been sidestepped in policy.

After the creation of a whole-of-society radicalisation prevention landscape in 2011 (and the statutory enforcement of the Prevent duty in 2015), a significant shift occurred in the UK’s governance of vulnerability with the creation of secretive vulnerability support hubs. These hubs are collaborations between psychologists/psychiatrists and counterterrorism policing in three major British cities. The Southern hub (also known as PLAD: Prevent Liaison and Diversion) operates from SO15 counterterrorism policing in London; the Central hub (also known as Prevent-in-Place) operates from Counterterrorism Policing West Midlands in Birmingham; and the Northern hub operates from Counterterrorism Policing North-West in Manchester. While Prevent’s policy documents present a homogenous vulnerability applying across the population (which, in practice, gives way to racialised referrals at disproportionate levels compared to demographic ratios), the hubs were an important step away from this (unconvincing) performance of a homogenous vulnerability within the population. As I will describe here, the hubs introduced ‘dividuated’ profiling, so that some subjects appeared more ‘vulnerable/risky’ than others on the basis of their mental health, their socio-economic profiles and situational factors. Effectively, the hubs created new subject positions around the risk factors identified in radicalisation guidance documents (such as “involvement in criminality, family influence, failure to integrate, or a sense of grievance or injustice” (NHS England 2017: 6)) – replacing a landscape of supposedly equally distributed, whole-of-population vulnerability with particularly risky subjects.

The hubs were created after a police-led study identified a higher-than-expected number of mentally unwell persons being referred to Prevent, as potential radicalisation risks (Dodd 2016). Rather than concluding that this statistic related to ableism and widespread discrimination against those with mental illness (causing them to be referred as ‘suspicious’ to Prevent), the Home Office equated ‘referral to Prevent’ with ‘potential terrorist conduct’. This shorthand logic went on to associate ‘mental disorder’ with potential terrorist conduct.

The Royal College of Psychiatrists has made it abundantly clear that psychiatry and deradicalisation are distinct practices – and while lone actor violence is more strongly associated (than group-based violence) with certain psychiatric disorders, it is nevertheless unclear whether treatment of any disorder would impact upon the attack planning (Royal College of Psychiatrists 2016). The Royal College has also cautioned against the use of risk assessment tools for terrorism and extremism, emphasising that the rare occurrence of terrorist violence (compared to non-political violence) makes prediction scientifically impossible (Royal College of Psychiatrists 2016). Despite these expert comments, the hubs understand that mental health treatment may reduce the terrorist risk posed by some referred to Prevent.

They each are funded to explore the significance of this ‘link’ (which is not really a link). The hubs utilise slightly different methodologies, relating to the approach of each cooperating NHS mental health trust and their assigned lead practitioner, but study the prevalence of psychiatric disorders in all those referred to Prevent in the northern, central and southern areas of Britain. Crucially, they establish a new dynamic in the discourse of ‘vulnerability to radicalisation’ – such that psychiatric cooperation with counterterrorism police creates ‘combined risk scores’ from medical data and police surveillance/information. Here, the stickiness of vulnerability and risk to a subject is ‘dividuated’ according to their racialisation as well as their individual behavioural flags and psychiatric diagnoses of note.²

To obtain the internal reports of the vulnerability support hubs, Medact (a London-based activist group) and their supporting authors engaged the Freedom of Information (FOI) request process at length. At first, the existence of the internal hub evaluation reports was denied by the recipients of the FOI request. Then, the documents were ‘found’, but their unredacted release to Medact was refused. After the intervention of the Information Commissioner’s Office, the reports were finally released to Medact – who published a full analysis of the contents (Aked et al. 2021) alongside the original documents.³ The reason for the obstruction was very apparent as soon as researchers opened the pages. Despite government assurances to the contrary, the hubs do not provide much that could be interpreted as mental health liaison pathways to people referred to Prevent. They rarely assess individuals or put them in touch with services. Rather, the hubs facilitate the transfer of confidential medical information to counterterrorism police by liaising with other NHS professionals. The vast majority of those referred to Prevent with diagnosable mental illnesses were, according to the internal reports, already receiving treatment or were known to mental health services from prior engagement.⁴ As such, the description of the hubs as a ‘liaison service’ providing pathways into mental health services is incomplete and imperfect.

Instead, the transfer of diagnostic information into the hubs (which are based within major counterterrorism policing centres) is used to administer a RAG rating – which combines medical information with police concerns to rank Prevent referrals as red, amber or green (hence: RAG) (Aked et al. 2021; see also Aked 2021). As the Royal College of Psychiatrists has confirmed, it is important that its members are cautious and resist the expectation from governments and security officials that psychiatric expertise can shed light on the terrorist risk posed by an individual (Royal College of Psychiatrists 2016).

And yet the vulnerability support hubs’ reports demonstrate that this is exactly how information about mental disorders is being used. The reports show that diagnoses of psychosis and schizophrenia are highlighted in reports about individuals of concern and are combined with police information (about the verbal abuse of minorities by these individuals or their recent suspicious movements on motorways) to issue a ‘red’ notice (Aked et al. 2021). Such a ‘red’ RAG warning results in hospital detention within the hour. For one case, involving schizophrenia and verbal abuse of ethnic minorities, the trigger for hospital detention by the hubs was a forthcoming right-wing demonstration in their area of residence. In another case (involving psychosis and motorway travel), the trigger for the hospital detention was the occurrence of the Westminster Bridge attack days previously. These contextual factors (an attack in previous days; a forthcoming right-wing demonstration) had no direct connection to the individuals – but were important in the decisions to section them under the Mental Health Act.

While threats to the safety of others can justify detention under the Mental Health Act, serious questions can be asked about the pre-emptive imagination of such threats here – given the chance interplay of slight changes in mental health, combined with police information about future demonstrations, which had little to do with the individual, and/or police fears about the potential for a copycat attack. The combination of police fears/information with diagnostic information is combined by the hubs, such that the interplay – and not any direct evidence that the individual plans to cause harm, alongside no evidence of serious deterioration in mental health – adds a securitising impetus to the assessment of the individual. The individual graded ‘red’ in such assessments is then detained against their will, for weeks or even months.⁵

Stopping short of hospitalisation, ‘amber’ gradings by the hubs are also considerably invasive. Evidence from the West Midlands hub demonstrates that – without the consent of the individual – the hub can request that local mental health teams monitor subjects for specific phrases or interests, and set up ‘tripwires’ for re-referral to Prevent,⁶ and the hub, if those phrases/interests come to light during treatment (Aked et al. 2021). This is a transgression of medical norms, whereby a person’s therapeutic treatment becomes a space of surveillance by the counterterrorism police (see also Rizq 2017).

The vulnerability support hubs mark a significant departure from the previous framing of vulnerability, evident in Prevent. The state is framed as vulnerable, in the securitised discourse of the hubs, not the individuals with serious disorders who might fall prey to radicalisers. Rather than conceiving of a (supposedly) equally distributed vulnerability across the population, the hubs protect the vulnerable state by associating individuals who have diagnoses of autism spectrum disorder, psychosis and schizophrenia with a profound, and sometimes immediate, threat to society that requires intervention. Despite being called vulnerable by the hubs, it is clear that individuals with disorders are being securitised as threats. It is the presumed vulnerability of the state that dominates the logic of the hubs, not those receiving mental health care. Ironically, the presence of a formal care need (mental illness; disability) in hub referrals escalates the state’s response beyond that of safeguarding. In standard safeguarding, a formal care need would open the possibility of a supportive safeguarding intervention – whereas through the vulnerability support hubs, it precludes a supportive multi-agency intervention and instead reaches for security measures, such as detention, to protect the vulnerable state.

A similar securitisation of vulnerability (where the signifier is no longer applied to an individual but characterises the state) has also recently occurred, through changes applied to the MAPPA system.

The Independent Reviewer of Terrorism Legislation and the redistribution of vulnerability and risk in MAPPA

In the aftermath of the Fishmongers’ Hall (2019) and Streatham (2020) attacks by recently released offenders, the UK Government commissioned the Independent Reviewer of Terrorism Legislation (IRTL) to review MAPPA for terrorist offenders. The recommendations made in the IRTL’s report have directly influenced provisions in the PCSC Bill 2021 (Home Office 2021; see also UK Parliament 2021: paragraphs 158–163), which became law in 2022.

The IRTL’s report on MAPPA contains a significant number of recommendations for the public protection arrangements. Among these, police powers to recall released offenders (particularly those thought to pose a risk of terrorism) to prison are recommended to be extended, and intelligence possessed by MI5 is recommended to be incorporated in MAPPA supervision of offenders thought to pose a terrorism risk (where such information exists). To facilitate this sharing of sensitive security information, the IRTL recommends that a representative of MI5 sits on relevant MAPPA case management meetings – directly integrating the intelligence services within certain probation procedures.

There are several civil liberties issues of concern, even before this chapter turns to the extension of severe restrictions to non-terrorist offenders. Firstly, the recommendations that extend police powers to recall offenders to prison include a deterioration in mental health. Where mental health is “associated with a risk of terrorism”, the report recommends that deteriorating mental health can be used as a pretext to recall offenders to prison for public protection reasons. The report states:

The justification for this recommendation is framed in terms of public protection and for removing the need for costly armed surveillance of an offender deemed to present a terrorism risk (IRTL 2020: 24). The deterioration in mental health is not framed as requiring a healthcare intervention, but a security intervention. The IRTL carefully explains that recall to prison on grounds of mental health deterioration is not a step to be taken lightly, but only because the offender’s right to appeal will require the authorities to present their reasons for recall. In this situation, the IRTL recommends that authorities tread carefully and prepare a ‘gist’ of intelligence that can be presented to the parole board – without disclosing the covert mechanisms by which this information was obtained. As such, the civil liberties concerns associated with recalling an offender on mental health grounds are circumvented (why not provide healthcare for the deterioration in mental health?), in favour of a description of protecting the intelligence services from the effects of disclosures.

Secondly, the IRTL’s report also recommends the involvement of MI5 on MAPPA case management where relevant. Prior to the PCSC Bill, terrorism risk assessment was undertaken through the ERG22+ scale – likely to be applied to offenders in a prison setting. The IRTL identifies an ‘over-reliance’ on the ERG22+ and OASys⁷ in MAPPA’s assessment of terrorism risk and instead recommends that MI5 is involved in active case management of released offenders (to share intelligence relating to the activities of family members) as well as the use of polygraph testing. Polygraph testing of terrorist offenders is recommended so that breaches of licence conditions, such as consuming alcohol or drugs, can be identified and used to justify recall to prison (IRTL 2020: 11). The IRTL notes the inadmissibility of polygraph test results in court but positively assesses the Government’s plans to use lie-detector tests to facilitate the recall of offenders for breaches of licensing conditions. Getting the person back into prison on a technicality is framed as appropriate and warranted. At all stages in the report, the prevalence of securitisation over civil liberties and a welfare-based response to vulnerability is evident – contextualised with little discussion of the acceptable boundaries of police activity in a liberal democratic state.

The IRTL is very critical of previous mechanisms for assessing the risk of violence and terrorism because they present an optimism bias that focuses on the protective factors that reduce risk, rather than explicitly tackling the potential imminence of the offender committing an attack (IRTL 2020: 6–7). As the application of an ERG22+ assessment would not have access to sensitive information possessed by the security services, and an ERG22+ assessment cannot take into account the effect of international ‘trigger’ events (such as the establishment of an ISIS caliphate), the IRTL recommends that security services are positioned on active case management through MAPPA – to aid the sharing of sensitive information (IRTL 2020: 7–9). In this way, the IRTL argues, the prediction of ‘terrorist risk’ can be made more dynamic and responsive. The presentation of terrorism risk assessment as having an optimism bias is directly indicative of a securitisation frame, which exposes those subject to MAPPA with increased risks of detention – even on technicalities, such as the potential impact of international events on the perception of their risk.

In all the recommendations, a greater securitisation of released offenders is evident – based on risk assessment of what they might do in future, the possible local effect of developments in international politics, attempts to discover licensing breaches through polygraph testing, and information possessed by the security services on a person’s – and their family’s – activities. There is no approved method for quantifying these factors and assessing their impact upon risks posed by terrorist offenders; instead, a reformed MAPPA will make subjective judgements under the powerful influence of members from MI5 and counterterrorism police forces. The calculation of these factors as risks (and the implementation of strict licensing conditions to protect the public) becomes even more concerning when the reformed MAPPA are applied to non-terrorist offenders, through the new category of the ‘terrorist risk offender’.

Terrorist risk offenders: the impact of MAPPA reform on vulnerability and securitisation

The Prevent strategy presented vulnerability to becoming terrorist as a population-wide landscape, where vulnerability could be identified through gut instinct of public servants – later to be measured and mapped through the application of risk assessment tools such as ERG22+, VAF and TRAP-18 (Terrorist Radicalization Assessment Protocol). This changed radically with the parallel structure of vulnerability support hubs and with the 2021 PCSC Bill proposals – including the IRTL’s recommendations for reforms. The IRTL recommendations introduce a new category of ‘terrorist risk offender’, which extends high-level MAPPA supervision to non-terrorist offenders upon their release – on the basis of a ‘perceived risk of radicalisation’. MAPPA management plans for offenders can include restrictions on residence upon release (including the requirement to live in ‘approved premises’ rather than return home), licence conditions that ban the individual from entering certain areas or premises, and even covert surveillance in extreme cases. As such, the proposed extension of MAPPA to terrorist risk offenders (who have not committed a terrorist offence, may not have even committed a violent offence, but are presumed radicalised) is a pressing civil liberties issue. As will be detailed here, the risk posed by terrorist risk offenders is not of verifiable nature, but is based on securitised suspicion in the aftermath of the Fishmongers’ Hall and Streatham attacks.

Previously, the definition of terrorist offender was self-explanatory. If an offender was convicted for breaching the Counterterrorism Act 2008, then they became a terrorist offender. Terrorist offenders are released into the community under Level 3 MAPPA arrangements. MAPPA Levels 2 and 3 both involve active participation by numerous local and national agencies. These offer more intensive case management than Level 1 MAPPA arrangements (which are reserved for registered sex offenders). Level 3 MAPPA case management involves the presence of senior officers who can authorise additional budgetary resources when required and respond to the potential for significant media coverage of the offender.

However the IRTL’s report on MAPPA reform, and the PCSC Bill 2021, introduce new terminology of the terrorist risk offender. They define this profile as:

The IRTL explicitly recommends that new powers are created to allow the police a warrant to search the premises of terrorist risk offenders to check their compliance with licensing conditions – just like terrorist offenders. These searches would be designed to discover, for example, “possession of a phone in violation of licence conditions” (IRTL 2020: 20; see also UK Parliament 2021: paragraph 160). The justification for searching non-terrorist offenders’ premises is based entirely on a presumed terrorist threat (such as being in contact with a terrorist network) posed by the non-terrorist offender. Similarly, the IRTL recommends that parole boards vary their practice so that urgent changes to licence conditions are accepted for terrorist-risk offenders (i.e. non-terrorists) (IRTL 2020: 21–22). The IRTL specifies that exclusion from particular areas is such a change that should be considered, given that more specific licence conditions (such as: “you must not communicate with person X”) would risk revealing intelligence to the offender. The IRTL gives the justification of preventing copycat attacks for this adaptation of parole board practice.

A great many of the existing MAPPA powers for terrorist offenders (such as requiring the released offender not to return home but instead live in approved premises) will become applicable to terrorist risk offenders. But terrorist risk offenders have never been convicted of a terrorist offence, so the application of a restrictive regime of licensing conditions becomes a civil liberties issue. In effect, the changes to MAPPA (proposed in the PCSC Bill, paragraph 163) replicate several components of terrorism prevention investigation measures) – which also restrict the civil liberties of persons in the community, where sufficient evidence is not available to prosecute.

Interestingly, while increasing the securitisation of released offenders thought to hold radical views, the IRTL also downgraded the presumption of security risk associated with highly vulnerable people, in certain situations. There is a notable ‘pulling in different directions’ occurring, for different communities of vulnerable people. The IRTL’s MAPPA reform report, and a speech the reviewer gave on its recommendations, draw a contrast between the limited level of MAPPA supervision required to manage ‘lonely’, often autistic terrorism offenders (the vulnerable) and the conversely high level of supervision required by those feared to have radicalised/converted in prison, or whose family has attracted MI5 surveillance. Here, the previous policy landscape of ‘undifferentiated vulnerability in the population’ is radically rethought, with greater securitisation of released offenders (who have not committed a terrorist offence) and the attribution of innocence and low risk to some offenders who have. How can this be?

Racialising terrorist risk: the innocence of the convicted (White) autist

In the pre-emptive discourse of ‘terrorist risk offenders’, it becomes especially important to pay attention to the racialisation encoded within both ‘innocence’ and ‘risk’. Applying licensing restrictions to non-terrorist offenders on the grounds of subjectively calculated risk alone (such as information that a non-violent offender has shown radicalised tendencies while incarcerated) introduces a dangerous potential for discriminatory decision-making.

The IRTL’s report on MAPPA reform, and subsequent speech to the Forensic Network Scotland in June 2021, both introduce a profound distinction between naïve and innocent terrorist offenders and their counterparts who pose a substantial risk of harm to the public. Returning to the IRTL’s comments on the limits of OASys and ERG22+ within MAPPA, they situate the innocence/risk binary not in the protective factors that can be fostered with offenders – but in the ‘imminence’ of their risk to the public. As such, one can be a convicted terrorist offender (perhaps the offence was downloading banned images) but be framed as very low risk, but the IRTL also conceives of non-terrorist offenders who must be placed under the highest levels of MAPPA supervision because of a perceived risk to the public. How is this possible?

The IRTL spends significant time disentangling the categories of ‘terrorist offence’ and ‘risk’ to support this perspective, expressing significant disappointment that current MAPPA treat all terrorist offenders as presenting the same level of risk – given their offences. The IRTL would prefer if the risk to the public was disaggregated from the index offence, such that – in some cases – non-terrorist offenders (potentially convicted of a non-violent offence such as fraud) could be treated as a higher threat than some terrorist offenders.

Primarily, the IRTL bases this argument (and the recommendations for MAPPA reform, now formulated within the PCSC Bill 2021) on the ‘innocence’ of many autistic or mentally ill terrorist offenders. In the MAPPA reform report, the IRTL states:

The report goes on to highlight the emerging profile of terrorist risk offenders as “lonely, vulnerable, self-radicalised individuals who are drawn to extreme views, usually encountered and reinforced online, many with poor mental health. These offenders are often convicted of possession of prohibited terrorist material and it is often difficult to distinguish the extent to which this is a type of obsession or a deeper attachment to a particular cause that could lead to acts of violence” (IRTL 2020: 30).

In case the allusions to fixation and obsession are not clear, the IRTL more directly specified, in the 2021 speech in Scotland, that the report meant offenders on the autistic spectrum. Speaking to forensic health practitioners, the IRTL identified an “extraordinary rise in the phenomenon of young men, frequently suffering from poor mental health or on the autistic spectrum, who are drawn to expressions of support for violent ideologies” (IRTL 2021: 1; see also IRTL 2020: 30). Immediately, when outlining this ‘emerging profile’ of neurodiverse individuals and those with poor mental health, the IRTL appeared to racialise the category as White, saying:

The IRTL went on to say that not all such offenders were linked to right-wing causes, but still, the association was powerful enough to make this generalised statement. This implicit racialisation of mentally ill and autistic offenders as White is important, because the IRTL frames such terrorist offenders as often being a low risk to the public. The IRTL even goes so far as to frame them as naïve and lonely, requiring social support to occupy their time rather than the harsher licensing conditions applied to others. Framing the obsessive pursuit of terrorist-linked content online as the behaviour of lonely, harmless people, and then racialising them as White by linking them to right-wing extremism, sets up an extremely problematic binary with those who are not deemed so innocent.

For the “lonely, vulnerable, self-radicalised individuals who are drawn to extreme views” (IRTL 2020: 30), the IRTL recommends that MAPPA gain access to resources and structures currently associated with Prevent – such as assessment by vulnerability support hubs. Here those with poor mental health and autistic spectrum disorder might receive care that mitigates their risk of committing another terrorist offence.⁸ Compare this with the IRTL’s extension of MAPPA supervision, and the recommendations made for new police powers to deal with non-terrorist offenders who are simply suspected of radicalisation.

A powerful contrast between ‘innocent and lonely’ offenders against those who pose ‘a high risk to the public’ is created. Unfortunately, one is predominantly (but implicitly) racialised as White through the association with right-wing extremism, whereas the remainder are implicitly racialised as Black, Brown and Muslim – given that the MAPPA reform report was directly provoked by the attacks of Usman Khan and Sudesh Amman. If the actions of these racialised Muslims provoked the report, then we can consider most of the recommendations are implicitly intended to deal with the danger posed by other such racialised Muslims; except where the reviewer offers a new categorisation of the ‘lonely, vulnerable and obsessive’ offenders, linked to the ideology and iconography of the far right. Promoting such a disparity in treatment, after release from prison, is unacceptable and unbecoming of any state – let alone one that purports to provide equal opportunities for all.

Despite its distastefulness, the racialisation of dangerousness/vulnerability in the MAPPA reform report is reminiscent of the first UK Prevent strategy, when Prevent funding was allocated according to the size of Muslim communities within local authorities. A large literature then criticised the Prevent strategy for creating a ‘suspect community’ of racialised British Muslims (Awan 2012; Breen-Smyth 2014; Kundnani 2009), a topic that also dominated civil society resistance to the first Prevent. The explicit racism of early Prevent was then addressed (or whitewashed) by the Prevent review of 2011 – which unlinked Prevent funding (from national government to local areas) from demographics.⁹ And yet, the IRTL’s recent work (2020; 2021) recovers the explicit racialisation of terrorist risk, and terrorist innocence, attempting to once again deploy it in statutory form – through contributing to the PSCS Bill 2021 (Home Office 2021; UK Parliament 2021).

Putting MI5 into psychiatric hospitals

This section explores some of the most extreme securitisation of vulnerable people in the MAPPA reform recommendations, where the actual vulnerability of psychiatric in-patients is disregarded – in favour of the state’s claim to vulnerability. It is proposed that MI5 will be directly integrated within the management of psychiatric in-patient care, where there is a suspicion of radicalisation. In these recommendations, the most vulnerable in society are reframed as the most dangerous – necessitating the extreme step of having intelligence services involved with psychiatric hospital caseloads, to protect the vulnerable state.

The IRTL report into MAPPA reform contains an annex specifically dedicated to terrorist risk offenders (i.e. non-terrorist offenders) who are detained in a hospital setting under the Mental Health Act. The consideration given by the IRTL extends to offenders who are detained in hospital in accordance with a court decision, or those serving sentences in prison who are later transferred to hospital. In all cases, these people are extremely unwell to the extent that medical professionals take over their care.

Prior to the implementation of PCSC Bill 2021, the IRTL was highly critical of arrangements for the sharing of information regarding the terrorist risk posed by those detained in hospital. The Joint Extremism Unit, together with counterterrorism police, share information regarding terrorist risk with the Mental Health Casework section (Ministry of Justice), so that any leaves of absence from hospital or transfers between hospitals can be accommodated without a gap in information sharing. However, the IRTL strongly criticises this practice, because no member of staff in the Mental Health Casework section has a sufficient level of security clearance to be appropriately briefed on the most sensitive security information (IRTL 2020: 61). As such, sensitive security information has to be ‘gisted’ by the Joint Extremism Unit and counterterrorism police, rather than explicitly informing decisions made by the Secretary of State (through the Mental Health Casework section).

The IRTL identifies that this set-up potentially compromises decisions taken on the offender’s location within the hospital, their conditions of detention, their access to leave of absence, considerations regarding discharge and decisions following release (IRTL 2020: 62). As such the IRTL recommends that MAPPA (including the presence of counterterrorism police representatives and MI5, where necessary) are extended into the hospital, so that pre-release planning, sensitive information sharing and patient management can take place from the moment of admission:

Furthermore, the IRTL recommends that those terrorist offenders (or terrorist-connected offenders) who take temporary leave outside psychiatric hospitals are not relieved of their notification requirements to police. That means that a seriously mentally ill patient enjoying a period of temporary leave from hospital is placed under the burden of registering their address with the police or face breaching those legal requirements. Is this a reasonable requirement that someone enjoying a temporary visit home from a psychiatric hospital could perform?

Problematically, this recommended cooperation between intelligence and security officials with psychiatric professionals, through reformed MAPPA structures, brings the intelligence services into the hospital – allowing them to influence the location and conditions of someone’s treatment in a healthcare facility. These recommendations, if accepted and implemented, will pose a significant risk to the separation of healthcare and security services in liberal society and could compromise the independence of psychiatric professionals in pursuing a treatment goal (for further reading, see: Heath-Kelly 2021). It is also an extreme securitisation of some of the most vulnerable people in society – in-patients of psychiatric facilities. Their vulnerability is silenced by this discourse, which instead frames the state as the most vulnerable actor in need of protection.

Conclusion: “It’s the economy, stupid”

This chapter has documented an ‘elastic band’ effect, where counterterrorism initiatives have radically altered the framing of who is vulnerable to radicalisation and whether their vulnerability results in a multi-agency welfare-based response – or a securitised intervention (where, paradoxically, the state is considered vulnerable). In the vulnerability support hubs, perversely (from a traditional safeguarding perspective), having formal care needs pivoted the state’s response to individuals away from multi-agency protection and towards securitised measures of surveillance and threat assessment. Yet in the MAPPA reform recommendations, autistic spectrum disorder was framed as a vulnerability that presented a ‘false positive’ flag for terrorist threat and was designated as often requiring less securitised interventions upon offenders. The measures seen in the hubs and in MAPPA reform are directly opposed in how they conceive autistic spectrum disorder and the risk posed to the state.

To understand this waxing and waning of vulnerability and securitisation framings in UK public protection discourse, we need to focus on the economic costs of preventing terrorism. The 2015 Counterterrorism and Security Act (building on the 2011 Prevent strategy) placed statutory counter-radicalisation duties on all public service providers – generating massive increases in referrals to Prevent and the associated costs for multi-agency panels in assessing those individuals and providing support packages. While the financial costs of support packages are not made public, the introduction of the duty led to referrals skyrocketing from 748 in 2012/13 (ACPO 2014) to 7,631 in 2015/16 (Home Office 2017).¹⁰ The year-upon-year figures have since hovered around 7,000 annual Prevent referrals (until the impact of the pandemic was felt in statistics for 2020/21). Comparing this stable annual rate to pre-duty levels demonstrates a staggering increase. The combined total of referrals across the years 2006–2013 number 2,653 (ACPO 2014) – a fraction of the yearly totals in the period after the statutory duty was introduced.

The increase in referrals contributes directly to the time and money spent on resolving them. As such, it is possible to understand recent reframings of ‘vulnerability’ as an attempt at increasing efficiency. Indeed, the IRTL comments on the need to increase efficiency in the recommendations for MAPPA reform, stating:

The vulnerability support hubs can also be interpreted as an attempt to render the Prevent system more efficient. By seeking to identify the particularly vulnerable/risky few, from a landscape of whole-of-society referrals, the resources of the state can be – it is imagined – better distributed by focusing intensely on individuals with mental health conditions and disabilities. These vulnerable individuals with care needs are suddenly reframed, through economic logics of efficiency, as potential threats to the state. The MAPPA reform report enacts the same logic of cost savings, but identifies the most threatening individuals as those thought to be radicalised in prison. The vulnerability of both groups (in terms of exposure to ‘extremist groomers’) is suddenly inverted to frame them as threats – to a vulnerable state.

By flexing the identification of vulnerability and threat, these reforms to Prevent and public protection can be seen as correctives to the 2011 Prevent strategy (where Prevent responsibility was relocated from Muslim local areas to all public service providers in the nation). The implementation of counterterrorism through the whole-of-society-vulnerability paradigm produced too many referrals, many of them irrelevant referrals (to paraphrase the IRTL) that did not correspond with those people planning attacks. As such, the landscape of Prevent – which framed the whole population as (supposedly) equally vulnerable to radicalisation – is now being supplemented with securitised correctives, which identify some mental disorders and disabilities as indicators of greater terrorist threat. Ironically, in a true safeguarding approach, the presence of these conditions would strengthen – not bypass – the need for a supportive, welfare-based intervention. Instead, UK public protection is responding by blurring the competencies of the intelligence services, counterterrorism police, MAPPA and mental health professionals to enact the faster detention, or increased surveillance, of those with specific care needs. Except where the IRTL provides specific exceptions for autistic terrorist offenders associated with right-wing extremism – for whom vulnerability is reclaimed as a state of innocence.