Bengaluru: Experts from across disciplines have called for a fundamental rethink of how digital tools and artificial intelligence are designed and deployed in autism care, urging that technology remain flexible, human-centred and responsive to individual needs. The recommendations emerged at the third Global Autism Convention organised by St John’s National Academy of Health Sciences, in association with several organisations.Speaking at a symposium titled Digital Tools & AI-based Interventions Friday, autism advocate and app developer Alice Mamaga Akosua Amoako from Ghana underlined technology’s potential to bridge global gaps in care while warning against a one-size-fits-all approach. “Some children are highly tech-savvy, while others may avoid screens entirely. Parents, caregivers, teachers, and professionals must understand each child’s individual needs, interests, age, and developmental level before choosing any digital or AI-based intervention,” she said.
Echoing the call for customisation, Matthew K Belmonte of Nottingham Trent University pointed out minimally speaking autistic individuals are often excluded from research despite having relatively intact comprehension. Presenting a Bengaluru-developed, open-source, iPad-mediated therapy, he stressed that digital platforms should function as enablers rather than constraints.He highlighted how alternative communication pathways can uncover abilities overlooked by conventional assessments. Citing an example, he said iPads typically assess responses based on the point of contact made by a child, whereas recognising the moment of departure from contact could be more meaningful. “When the departure is recognised, the computer’s response can be more congruent to the user’s intention,” he explained.Bringing an engineering and neuroscience perspective, Suresh Sundaram, professor at the Indian Institute of Science (IISc), challenged rigid behavioural models. “We all recognise an apple even when we can’t explain how—autism works the same way. Behaviour emerges from functional brain networks, muscle memory, and subtle emotional signals, not deterministic rules,” he said, advocating probabilistic, data-driven approaches using multimodal data such as brain signals, behaviour, language, vision and physiology.He noted stereotypical behaviours in autistic children can appear suddenly without obvious triggers, and that identifying underlying causes could help interrupt escalation. “By using vision-based systems and sensors, it becomes possible to study gaze patterns and emotional changes more closely,” he said. “These micro-emotions play a critical role in behaviour and can easily go unnoticed. Children are highly sensitive to such cues and often read emotional changes more accurately than adults.“The five-day convention, themed ‘Autism in the Global South: Culture, Community, and Neurodiversity’, was inaugurated Wednesday. Dr Vijaya Raman, professor of psychiatry at St John’s Medical College Hospital and chair of the convention, said the focus was deliberately grounded in local realities.“This year’s convention was designed specifically for low-resource settings like ours, where solutions from the West are often not directly replicable. We wanted to bring parents, clinicians, therapists, educators, and researchers together to look at autism through the lens of culture, community, and neurodiversity, and to develop approaches that work within our realities—strengthening families while reducing caregiver burden,” Dr Vijaya said and added that digital tools and AI are also transforming early identification – flagging signs of autism much earlier than clinical observation alone — sometimes as early as 9 to 12 months.
