Autism, AI and the workplace: why different minds belong in the field

Simon Todd in 26 Oct 2026

In September 2025, I was diagnosed with Autism Spectrum Disorder (ASD). I had already completed my Master’s degree. My dissertation looked at how artificial intelligence could support autism diagnosis and therapy, so autism and AI were part of my studies before I understood how closely the subject connected to my own life.

By the time I was diagnosed, I had already established a career in technology. For me, that is a reminder that ASD need not be a barrier to success. I continue to take on new work and learn as the field changes. I know my experience is only one person’s experience. Autistic people have different strengths, needs and ambitions. An ASD diagnosis should not be treated as a reason someone cannot succeed in technology or AI.

Autism in my working life

Today, I work in AI, architecting AI solutions. I also look at how AI may shape the future and explore innovative ways it could be used. One example I have written about is using cameras and edge AI on bin lorries to spot road deterioration as they follow their usual routes. The vehicles would carry on collecting waste while helping councils build a clearer picture of how local roads are changing.

Music (Aastralis Music) is one place where I notice the way my mind works. I enjoy writing lyrics and developing musical ideas. I can stay with a line or melody until it feels right. Mastering a song is a different matter. The technical detail and number of small decisions can overwhelm me. Sometimes my brain simply cannot manage that stage.

I can enjoy and do well at one part of making music, then struggle with another. The same is true at work. Autism does not give me a uniform set of strengths. It shapes how I approach things, and the effect depends on the task and the environment.

People often associate autism with logical thinking, attention to detail and intense concentration. Some autistic people describe these as strengths, and research has recorded strengths reported by autistic workers (Cope and Remington, 2022). Reviews also caution against assuming that every autistic person shares these traits or that they point to one particular career (Griffiths et al., 2024).

Do you need to be autistic to work in AI?

No. AI needs people with many different skills and backgrounds. It involves research and engineering, but also product design, data quality, infrastructure, security, governance, communications and knowledge of the sectors where AI is used. A diagnosis should never become an informal job requirement, and autistic people should not be channelled into technical roles because of assumptions about how they think.

The research directly connecting autistic people with AI work is still limited. A letter in the journal Autism asks whether AI will create opportunities or add to barriers autistic people already face in employment. Early research with autistic software engineers looks at their experiences, strengths and work preferences. Another study explores how autistic workers use large language models for workplace communication, including concerns about the quality of advice an AI tool may give. Research into visual data annotation considers whether this work might offer an employment pathway connected to AI development.

A recent Fortune article has brought this debate into the boardroom. Palantir CEO Alex Karp, who is dyslexic, argues that neurodivergent people could be well placed as AI changes work. He is talking about neurodivergence broadly, including autism, ADHD and dyslexia. The article also includes autistic psychologist Daniel Wendler, who sees value in different perspectives but warns against describing every neurodivergent person as having a special “superpower.” That kind of stereotype can leave out people whose strengths are less visible and can hide the real challenges they face.

That balance matters to me. Different ways of thinking can add value, but a diagnosis is not proof of a particular talent. Employers need to recognise the individual, give people a fair chance to show what they can do, and provide support that fits their needs.

AI can support people, with care

My Master’s dissertation explored how AI might help with autism diagnosis and therapy. I was interested in whether technology could help people get useful support sooner and make that support more personal.

There is now a growing body of research on AI for autism screening, assessment and therapy. Reviews describe promising approaches, while also pointing to weaknesses in study quality, validation and representation (Reilly et al., 2025; Akemoğlu et al., 2026). These systems should support clinicians and autistic people as they make decisions. They should not be treated as a diagnosis by themselves.

AI tools may also help with everyday work, for example by making written communication clearer, organising information or breaking a complicated assignment into smaller steps. Wendler raises some of these possibilities in the Fortune article. The tools need to be tested with the people who use them, and workers should be able to decide whether they help. If a system assumes that everyone communicates in the same way, it can misunderstand autistic people and repeat familiar workplace problems. AI should be designed with autistic people involved from the start, and people should be able to question its suggestions.

The workplace matters

Tools alone will not make a workplace inclusive. Clear instructions, predictable processes, written communication and attention to sensory needs can all help people do their jobs. Research based on autistic workers’ accounts describes barriers around recruitment, workplace communication, sensory demands and whether a role fits the person (Southey et al., 2024).

Employers should ask what helps each person do their work well. They should not rely on stereotypes about autistic “superpowers” or assume that one set of adjustments will work for everyone. The UK’s Buckland Review of Autism Employment and the Expert Academic Panel on Neurodiversity both point to the need for practical, individual support and more inclusive workplaces.

Neurodiversity and human evolution

I believe neurodiversity may be part of human evolution. Different ways of noticing and solving problems may have helped human groups adapt. Researchers are examining evolutionary explanations for autism, including how certain traits may have contributed to different roles in human societies (Hunt and Jaeggi, 2025). This remains an area of debate, not a settled account of why autism exists.

Evolution does not make every trait useful in every situation, and some autistic people need substantial support. My belief is that human variation has value, while the experience of disability and difficulty also deserves to be taken seriously.

Making room for different minds

My diagnosis came after my studies and years after the start of my career. It helped me understand myself, but it did not decide what I could achieve. I write music and work in AI. I also struggle with some tasks, including mastering a song as well as sudden changes to my schedule.

I would like the AI industry to be a place where autistic people can contribute if they want to, and where they can get the support that works for them. That benefits the people who work in the field and the people who use the systems it creates. No one should need an autism diagnosis to belong in AI, and no one should be ruled out because they have one.

Having collated a lot of references from my Masters, I like to also keep reading articles and studies, below are some of those.

itsabout.ai