Fresh official data from the UK Home Office has reignited serious concerns about the discriminatory impact of the country’s divisive Prevent counter-extremism strategy, after revealing that nearly one in five people referred to the programme between October 2024 and September 2025 had recorded autism diagnoses or suspected cases.
An independent analysis of the newly released dataset shows that 1,833 of the total 9,957 referrals logged over the 12-month period were linked to autism: 1,199 were confirmed diagnoses, while an additional 634 were marked as suspected cases. The annual referral total itself marks a historic high, the highest recorded since the programme began tracking national data in 2015, representing a 39% jump from the previous year’s figures for England and Wales. Overall, 36% of all people referred to Prevent during the reporting period had at least one recorded mental health or neurodivergent condition, with autism the most frequently documented neurodiverse trait across all cases.
Jacob Smith, a policy and advocacy officer at human rights group Rights and Security International, told Middle East Eye that the sharp rise in autistic people entering the Prevent system is “particularly concerning” and lays bare long-standing structural flaws embedded in the government’s counter-extremism framework. “We’ve known for years that autistic people are overrepresented in Prevent referrals, but the scale of this increase over the last 12 months is especially worrying,” Smith explained. He added that Prevent has faced accusations of systemic bias and discrimination since its launch: the programme initially disproportionately targeted British Muslim communities, and while demographic patterns have shifted over time, Black and Asian people are still vastly overrepresented in referral data, a trend that has now spread to neurodivergent groups including autistic people.
Alongside the autism findings, the data revealed a 20% year-on-year rise in referrals linked to far-right extremism. Demographic breakdowns confirm that Asian people account for 18% of referrals where ethnicity is recorded – twice their share of the general population of England and Wales. Black people are also overrepresented, making up 8% of referrals compared to just 4% of the overall population. Despite repeated government warnings that “Islamist extremism” remains the UK’s top domestic security threat, referrals categorized under this label have fallen to just 8% of all active Prevent cases.
Smith argued that the overrepresentation of marginalized groups stems directly from the programme’s vague, subjective referral structure. Public sector workers across the UK are legally required to refer any individual to Prevent if they are deemed to pose a potential terrorism risk, a mandate that leaves huge room for implicit bias to shape decisions. “Prevent is designed in an extremely vague way that leans almost entirely on referrers’ gut instinct,” he said. “This effectively encourages bias, whether intentional or unconscious, to creep into decision-making. That, to me, is the core reason we consistently see certain groups overrepresented in the system.”
The Home Office data also shows that the share of cases involving autism actually rises as cases move deeper into the Prevent process. Roughly 23% of cases reviewed by the multi-agency Channel panels – the bodies that assess radicalisation risk and approve de-radicalisation support – involved recorded autism, and that figure climbs to 24% for cases formally accepted into the government’s de-radicalisation programme.
These statistics are the second set of official Prevent data to include mental health and neurodiversity tracking, following the rollout of the new Prevent Case Management Tracker (PCMT) system in 2024. The Home Office acknowledged that the new system has improved how conditions are recorded, but cautioned that data quality is still evolving. The department also stressed that the figures do not prove any causal link between neurodivergence and susceptibility to radicalisation. “It is important to note that these findings describe associations within the data rather than causal relationships,” the official report reads. “The analysis does not indicate that being neurodiverse or having a particular mental health condition increases or decreases the likelihood of progressing through the Prevent or Channel process.”
Even so, Smith warned that the current system actively reinforces harmful, criminalizing stereotypes about autistic people. “A lot of the framing used by Prevent points to core autistic traits – for example, hyperfixation – and implies that these characteristics make autistic people more likely to pose a security risk,” he explained. “Instead of addressing the underlying bias in the system, they are doubling down by using autism diagnoses themselves as an explanation for overrepresentation.”
Under Prevent’s operating structure, after an initial screening and assessment, referrals deemed to carry a risk of radicalisation are passed to a multi-agency Channel panel, chaired by local authorities. These panels are tasked with evaluating an individual’s risk and deciding whether to approve a tailored support package to address perceived radicalisation. Of the 1,100 Channel cases accepted in the year to September 2025 with recorded ethnicity, 76% were white, 14% were Asian, 5% were Black, and 4% were categorized as other ethnicity.
Crucially, Smith added, even referrals that do not result in any further action can leave people facing long-term negative consequences. “As we documented in our 2024 report *Caught in the Web*, information about Prevent referrals – including cases that are dropped, found to be mistaken, or made with malicious intent – is shared widely across policing bodies, security services, immigration authorities and other public sector agencies,” he said. “While much of Prevent’s work operates behind closed doors, there is a very real risk of lasting harm for people who are referred, affecting not just their interactions with police but their access to all kinds of public services.”