A tool can look clever and still make a day harder. This review is not asking anyone to fear technology. It is a quiet reminder to look for evidence, choice and a real benefit in everyday life.
- Only two of the 13 studies met the review’s quality standards without reservations.
- A tool can support a person without replacing their choices or the people around them.
- If it adds pressure or does not help, stopping is an option.
Questions you can take with you
You do not need to become an expert. These questions can help you talk with a professional or support provider.
- What is this tool meant to help with, exactly?
- Who sees or keeps the data?
- Can I pause, change settings or say no without losing support?
Want the study details?Sample, method and evidence type
- Studies reviewed
- 13
- Met quality standards
- 2
- Common use
- Supporting adults, not replacing them
Read on for what was studied, what was found and what remains uncertain.
What this review did
The authors gathered 13 studies of interventions that used AI in some way. They covered very different things: communication, participation, emotion recognition, play and more. In most cases, the technology was meant to support a parent, teacher or clinician rather than work alone.
Many papers reported a positive short-term change. But positive is not the same as proven. The studies used different designs and outcomes, and the reviewers found weak experimental control, incomplete reporting and small or limited data in many of them.
Why ‘personalised’ is not enough
A digital tool can be well made, pleasant to use and still not have good evidence behind it. It can also fit one person and feel tiring, intrusive or simply unhelpful for another. The study does not prove that AI is harmful or useless. It says that the research base is still early.
A useful question is not ‘is this AI?’ but ‘does this help this person, in this setting, without taking away their choices or adding pressure?’ The answer deserves more than a product claim or a nice demo.
Keeping the human part in view
The review does not support replacing human judgement with an app or a camera. It also does not identify one tool that families should buy. Privacy, data use, accessibility and the person’s own comfort still matter, even when a tool has been studied.
If a digital option is being considered, it can help to make the goal very small and clear, then notice whether life gets easier, harder or no different. This is not medical advice. It is permission to ask for evidence and to stop if a tool is not a good fit.
Limitations to keep in view
- The review includes only 13 studies, with different tools, participants and outcomes.
- Its findings cannot tell us whether a specific commercial product works, is safe or protects privacy.
- Most reported effects were short-term, and the review found substantial limits in study quality.
A Systematic Review and Quality Appraisal of AI-Augmented Intervention Studies for Autistic Individuals
Akemoğlu Y, Roberts EB, Xiong J
Autism · 2026
This article provides general information and does not replace individualized medical, psychological or educational advice.