Project funded by the European Artificial Intelligence & Society Fund
Artificial Intelligence and the Rights of Autistic People (AIRA)
Developing new knowledge, training, and advocacy resources
AIRA is a two-year project (2023-2024) that maps the uses of artificial intelligence affecting the lives of autistic people, and builds up the capacity of autistic people to advocate for themselves.
AIRA focuses especially on the uses of artificial intelligence in diagnosis, treatments, interventions and augmentative & alternative communication for autism. These are mapped in the project’s literature review, initiated in early 2023. However, other more general uses of AI involve risks we share with various different minorities and disability groups, and the project will address these, as well.
Involvement of autistic people in AI research and product development have been minimal. AIRA maps autistic people’s knowledge, perceptions and needs regarding different uses of AI, with the aim of helping autistic voices to be heard in this field. The preparation of a multi-language survey for this was completed in 2023, in collaboration with Professor Ilse Noens from the Belgian university KU Leuven. The results will be used in the next stages of the project, and a peer-reviewed article is expected to appear in 2024.
The ultimate goal of AIRA is to develop the capacity of EUCAP member organisations and the wider autistic community in Europe to advocate for autistic people in matters related to AI, supporting them to influence legislation, policies, research and industry. Training and materials for advocates will be made available in 2024.
Literature on AI and autism has been mapped by three autistic researchers from three countries: Khiah Strachan from Scotland, Imke Heuer from Germany and Kat Van der Poorten from Belgium.
“If AI is being developed for autistic people, to make it as ethical as possible, we should be involved.”
In their preliminary report, the team noted that over the past eight years there has been an exponential increase in the number of articles focusing on AI-based approaches to autism. They identified a total of 1070 articles, grouped into four libraries based on the main themes. Diagnosis was by far the most common theme, with over 500 articles.
Overall, research involving AI and autism is largely focused on the application of machine learning algorithms to identify biomarkers and brain features associated with autism, as well as the development of AI-based diagnostic tools with the intent to facilitate early detection and intervention.
Among the 139 articles on interventions that the team identified, the emphasis was on robot assisted therapy, wearable AI technology, and AI assisted mobile applications. The team’s impression was that the majority of AI interventions included in this review focused on “lessening autistic behaviours” or “curing” autistic people rather than improving quality of life.
“I feel AI will play a crucial role in respect to the practice of diagnosing autism, to interventions, to research in general, but also in respect to policy issues such as access to health care. This is going to have a huge impact on autistic persons — and thus, the autistic community needs to have a voice in this field.”
Biomedical research involving AI and autism should follow ethical guidelines and regulatory frameworks to ensure the ethical and responsible use of data, protect the rights and dignity of autistic individuals, and minimise the potential harms associated with data breaches or misuses. Such guidelines were not obvious in the included articles, and may in some cases be lacking completely.
The researchers also commented on the lack of autistic people as research partners and co-creators of AI-based tools and interventions, as well as in the design and evaluation of research studies, and suggested that the autistic community could provide valuable feedback on the usability, accessibility, and effectiveness of novel tools and approaches.
The team’s work will be prepared for publication in a peer-reviewed journal.
“Incorporating feedback from the autistic community into the design, development, and use of AI systems can lead to more inclusive, effective, and ethical AI technologies that benefit everyone. Autistic individuals can provide valuable insights into how AI systems impact us, as well as how to design AI systems that are accessible and accommodating to us.”
Kat Van der Poorten