Algorithmic Risk and Neuroinclusion

AI, Psychosocial Safety, and the Rights of Neurodivergent Workers in Aotearoa New Zealand

Abstract

Artificial intelligence (AI) is increasingly embedded in workplace management systems, reshaping how performance is measured, risk is assessed, and psychological safety is governed. For neurodivergent workers, including those with autism spectrum conditions and attention deficit hyperactivity disorder (ADHD), these developments present both significant risks and genuine opportunities. AI systems calibrated to neurotypical norms can intensify workplace surveillance, encode cognitive bias, and erode the reasonable accommodation obligations that underpin inclusive employment. Reasonable accommodation refers to workplace adjustments that allow a person with a disability to perform their role effectively. At the same time, well-designed assistive AI tools hold real potential to support communication, workload regulation, and early identification of psychosocial harm at an organisational level. This paper examines both dimensions and argues that Aotearoa New Zealand’s current regulatory framework is structurally unfit to govern either equitably. The Responsible AI Guidance for the Public Service (Department of Internal Affairs, 2022) remains non-binding and neurotypically framed. The Health and Safety at Work Act 2015 imposes no AI-specific psychosocial duties. International instruments, including the United Nations Convention on the Rights of Persons with Disabilities and the European Union AI Act (2024), offer stronger benchmarks that New Zealand has not translated into enforceable domestic protections. Te Tiriti o Waitangi obligations apply across this space but are absent from every current instrument examined. The paper proposes reforms to align AI governance with neuroinclusion principles, psychosocial safety law, and Te Tiriti-based obligations.