The Impact of AI in the Legal System on the Neurodivergent Community
What does ‘suspicious behaviour’ look like to you? For an AI algorithm, it might just be a pause that is too long, an answer that is too detailed or a gaze that does not quite meet expectation. However, for a neurodivergent individuals, these behaviours may be entirely ordinary. As artificial intelligence is becoming increasingly integrated within decision-making, the line between difference and risk is no longer being drawn by human judgement alone, but by the systems trained to recognise patterns, patterns that not everyone fits.
AI in the legal system has been often framed as a solution to bias, validated by a 2025 UNSW research study that found more than 80 cases of AI used within Australian courts. By relying on data rather than human instinct, it has been thought to provide a more consistent and objective approach to decisions including bail and credibility assessments. Yet, this assumption of neutrality is exactly the crux of the issue. These AI models have been trained on historical precedent, which reflects existing norms about how individuals are ‘expected’ to behave. But for those who are neurodivergent, whose communication styles and behavioural patterns may deviate from these norms, this creates a structural disadvantage.
For example, behaviours associated with autism, including reduced eye contact and atypical tone or impulsivity can be misinterpreted within AI models as indicators of non-compliance or risk. Unlike human’s ability to discern, which can be questioned or challenged at present, algorithmic bias is becoming obscured within complex systems. As a result, while not always resulting in punishment, unequal outcomes can still be produced. This raises concerns over how credibility itself is constructed within legal contexts, as the law often relies on implicit expectations of consistency and expression that neurodivergent individuals may differ from. When AI is used to interpret such behaviours, there is a large risk that they reinforce a narrow standard of what is considered ‘reliable’, effectively excluding those who fall outside it.
It’s important to note that this does not argue against the use of AI in law, as these technologies continue to evolve and offer many benefits including efficiency. However, their usage must be critically examined, as without an awareness of how neurodiversity intersects with algorithmic criteria, AI may replicate, and even amplify, the same inequalities it seeks to eliminate. Thus, ensuring justice requires not only more relevant technology, but a more inclusive understanding of human difference.
Legg, M. (2025, September 29). More people are using AI in court, not a lawyer. It could cost you money – and your case. UNSW Sites. https://www.unsw.edu.au/newsroom/news/2025/09/more-people-are-using-ai-in-court-not-a-lawyer-it-could-cost-you-money-and-your-case
For any questions or feedback, please reach out at accessibility@utslss.com. All communication will be handled with confidentiality unless otherwise stated.