AI Eye-Tracking System Cuts Autism Diagnosis Wait Times, Dubai Research Shows
Researchers develop eye-tracking AI to accelerate specialist referrals for young children.
Researchers at the University of Dubai have built an AI screening tool that pairs eye-tracking technology with developmental assessments to flag children who may need specialist evaluation for autism spectrum disorder. The goal is direct: compress the gap between initial screening and formal diagnosis, a delay that currently stretches several years for many children.
The screening mechanism is straightforward. Children watch short videos containing social and geometric scenes while eye-tracking equipment records their visual attention patterns, including gaze direction, fixation duration and scanning behaviour. At the same time, parents complete a developmental milestone questionnaire. The AI model synthesises both data streams to produce an early risk indicator that can guide referral decisions.
Additional reference context is available at https://thesector.com.au/article/dubai-researchers-ai-tool-earlier-autism-screening-children.
What sets this approach apart is technical efficiency. Rather than converting eye-tracking data into images first, the system analyses raw tracking information directly. That design choice allows the software to run on tablets and smartphones without demanding high-performance computing infrastructure, a practical advantage for deployment across varied healthcare settings. The entire screening takes roughly two to three minutes and involves no invasive procedures.
Initial testing has yielded results the research team describes as encouraging. Benchmark accuracy reached around 96 per cent, though researchers are clear that the tool functions as a screening aid and referral support mechanism, not a diagnostic instrument. Clinical validation work continues across multiple healthcare environments to establish how the system performs under real-world conditions.
The timing problem it addresses is well documented. Autism can often be detected during a child’s first two years of life, yet many children do not receive a formal diagnosis until around age four. Earlier identification could accelerate access to intervention and support services, the researchers argue, because developmental interventions tend to show greater effectiveness when initiated sooner.
By contrast, established screening approaches rely primarily on parent-completed questionnaires, such as the Modified Checklist for Autism in Toddlers. By incorporating objectively measured eye movements alongside developmental information, the Dubai team is investigating whether AI can supply an additional data source to strengthen screening decisions.
The project sits within institutional and funding structures specific to Dubai’s research ecosystem. The Dubai Research, Development and Innovation Grant Initiative, established within the Dubai Future Foundation, provides financial support. Leadership comes from the University of Dubai’s College of Engineering and Information Technology, working in partnership with Emirates Health Services and Al Amal Psychiatric Hospital. Three universities with Australian connections also participate: the University of New South Wales, Macquarie University and the University of Wollongong in Dubai.
The implications reach beyond clinical settings. While autism screening and diagnosis formally sit within health and specialist assessment pathways, early identification developments carry relevance for early childhood education and care providers. Educators observe children’s communication, interactions, play and development over time and may contribute observations during conversations with families and other professionals. The emergence of AI-assisted screening tools therefore raises broader questions about how technology could complement existing developmental monitoring and referral processes already embedded in educational environments.
The researchers have been explicit about the tool’s limitations and intended scope. It is designed to assist screening and prioritise specialist referrals, not to replace clinical assessment or deliver an autism diagnosis. The system remains under clinical validation, meaning its effectiveness across different settings and populations must be established before any wider rollout. That cautious framing reflects a deliberate choice to verify performance across diverse contexts before implementation at scale.
Whether the tool’s accuracy holds across populations beyond its initial testing cohort is the question that clinical validation must now answer.
Q&A
What is the primary function of the AI screening tool developed by University of Dubai researchers?
The tool pairs eye-tracking technology with developmental assessments to flag children who may need specialist evaluation for autism spectrum disorder and to compress the gap between initial screening and formal diagnosis. It functions as a screening aid and referral support mechanism, not a diagnostic instrument.
What technical advantage does the system's design provide for deployment?
Rather than converting eye-tracking data into images first, the system analyses raw tracking information directly. This design choice allows the software to run on tablets and smartphones without demanding high-performance computing infrastructure, making it practical for deployment across varied healthcare settings.
What does clinical validation need to establish before wider rollout?
Clinical validation work must establish how the system performs under real-world conditions across multiple healthcare environments and whether the tool's accuracy holds across populations beyond its initial testing cohort.
What timing problem does the screening tool address?
Autism can often be detected during a child's first two years of life, yet many children do not receive a formal diagnosis until around age four. Earlier identification through the screening tool could accelerate access to intervention and support services, as developmental interventions show greater effectiveness when initiated sooner.