2025, Vol. 5, Issue 2, Part A
AI and assistive technology in autism rehabilitation: Opportunities and ethical challenges
Author(s): Elena Petrova
Abstract: Autism spectrum disorder (ASD) is a complex neurodevelopmental condition that requires timely and sustained intervention to improve developmental outcomes and quality of life. The rapid advancement of artificial intelligence (AI) and assistive technologies (AATs) has opened new pathways for early detection, personalized rehabilitation, and scalable therapeutic delivery. This study aimed to examine the opportunities and ethical challenges associated with AI-enabled AATs in autism rehabilitation, focusing on their clinical efficacy, stakeholder perspectives, and governance implications. A mixed-methods approach was employed, integrating quantitative synthesis of 42 published studies with qualitative thematic analysis of interviews conducted with 25 stakeholders, including clinicians, AI developers, and parents of autistic children. Quantitative data were extracted from peer-reviewed articles between 2015 and 2025, covering video-based diagnostic AI, eye-tracking biomarkers, speech/prosody classifiers, wearable biosensors, socially assistive robots, and virtual/augmented reality interventions. A random-effects meta-analysis was applied, yielding a pooled standardized mean difference (SMD) of 0.64 (95% CI: 0.51-0.78, p<0.001), signifying moderate-to-large improvements across domains of social communication, self-regulation, and adaptive skills. Subgroup analysis revealed greater efficacy in preschool children (SMD=0.78) compared to adolescents (SMD=0.52), highlighting the importance of early intervention. Wearable biosensors and VR/AR interventions produced the strongest gains, whereas video-based and robotic systems demonstrated moderate but consistent benefits. Qualitative findings identified four recurring ethical themes: privacy and data governance, algorithmic bias and fairness, user autonomy and consent, and implementation feasibility in resource-limited contexts. These concerns intersect with international regulatory frameworks such as the EU AI Act, Good Machine Learning Practice guidelines, and WHO’s ethics of AI in health. The study concludes that while AI-enabled AATs can substantially augment autism rehabilitation, their long-term success depends on embedding privacy, inclusivity, interpretability, and participatory co-design into their development and deployment. Practical recommendations include robust data protection, culturally diverse training datasets, clinician and caregiver training, equitable infrastructure investment, and continuous outcome monitoring. Collectively, these findings emphasize that AI can act as a catalyst for more accessible, individualized, and ethically sound autism rehabilitation when guided by human-centered values and governance.
Pages: 16-21 | Views: 409 | Downloads: 226
Download Full Article: Click Here
How to cite this article:
Elena Petrova. AI and assistive technology in autism rehabilitation: Opportunities and ethical challenges. International Journal of Autism. 2025; 5(2): 16-21.
Other Journals