One of the most complex questions practitioners face in cases of online extremism is deceptively simple: Did the person actually intend to act?
Findings from this research suggest that, particularly where neurodivergent people are involved, intent is often unclear, contested, or misunderstood. Algorithms and other online environments may facilitate escalation, but intent does not always develop in a linear or predictable way and treating it as if it does can increase both risk and injustice.
Algorithms as “trip wires”, not just pathways
Participants frequently described online algorithms as “trip wires” or “traps”, systems that pull individuals from casual engagement into more extreme spaces without a clear moment of decision‑making.
For those who are socially isolated, intensely interested in particular topics, or seeking stimulation, online spaces can offer belonging, structure, and identity. Algorithms reward attention, repetition, and escalation, regardless of whether engagement is playful, exploratory, or ideological.
What becomes difficult for practitioners is identifying when, or if, this engagement translates into criminal intent.
“It wasn’t real – it was online”
A recurring theme across interviews was the perceived distance between online behaviour and the offline world. Participants described neurodivergent individuals engaging in spaces that resembled fantasy, role‑play, or performance, often referred to as live action role play (LARP).
In these spaces, individuals could experiment with identity, language, and status in ways they could not offline. Online anonymity functioned as a form of masking, not unlike the masking many neurodivergent people describe in everyday neurotypical environments. Within this context, statements were often framed as:
- “just saying it for likes”
- “saying it because others expect it”
- or “not really meaning it”
This created significant uncertainty around mens rea (a Latin term used to describe intention, used in criminal settings to determine if a defendant had a culpable state of mind during a criminal act). Practitioners reported that many neurodivergent individuals insisted they would never take real‑world action, even after engaging in highly concerning online behaviour.
The risk of pathologising disengagement from “reality”
While this blurring between online and offline worlds is relevant, the way it was sometimes described raises important concerns. Some accounts risked portraying neurodivergent people as lacking an appropriate grasp of reality, rather than recognising that digital environments fundamentally reshape how meaning, identity, and consequence are experienced by many people.
When intent is assessed through a neurotypical lens, behaviours can be misinterpreted; either overestimated as ideological commitment or underestimated as harmless fantasy. Both perspectives can carry risk.
When intent starts to shift
Although many cases appeared to lack clear intent, participants also described situations where online engagement became unidirectional. Once certain thresholds were crossed, individuals could feel “trapped”.
In these cases, choice narrowed. If behaviour did not escalate, consequences followed. This is where some participants observed movement from online expression to offline planning or action, even if initial engagement had been exploratory rather than ideological.
Importantly, however, this process was not inevitable. In three interviews, practitioners described individuals who were successfully de‑escalated, redirected, and supported out of extremist risk pathways.
Reflections for practice
This finding highlights the need for careful, nuanced assessment:
- How do we distinguish between performance, fantasy, and intent?
- Are we equating online language with belief too quickly?
- Do we recognise how algorithms amplify behaviour without amplifying understanding?
- And do our systems allow space for disengagement, or only escalation?
Mens rea cannot be inferred solely from digital behaviour, particularly in environments designed to reward attention rather than meaning. Without that distinction, systems risk criminalising vulnerability while missing genuine ideological risk.