The human factor: how AI turns the spotlight on us

AI is powerful, pioneering, divisive and unknown. We ask two people working closely with the technology to share their insight into where AI belongs in humanity… 

Refik Anadol

Media artist and recipient of the Lau Fellowship in Creativity and AI, University of Oxford

“AI forces us to confront difficult truths. Who owns memory? Whose voices are amplified, and whose are missing?”

I believe the more urgent question may be how humanity chooses to belong within the age of AI. Artificial intelligence is not an external force arriving from elsewhere; it is a reflection of us. It carries our languages, our archives, our values, our blind spots. When we train an AI model, we are encoding collective memory. In this sense, AI mirrors humanity, often with uncomfortable clarity. 

For me, AI belongs in humanity where imagination, ethics, and care meet, and it belongs as a collaborator. When used responsibly, AI can help us perceive patterns that were previously invisible: in nature, in writing, in history, in emotion. It can reveal the scale of our interconnectedness across time, place, and culture. 

At the same time, AI forces us to confront difficult truths. Who owns memory? Whose voices are amplified, and whose are missing? What values are quietly embedded in systems that feel neutral but are not? These are not technical questions alone; they are deeply humanistic ones. This is why collaboration with the humanities is essential. When AI enters conversation with humanistic enquiry, it can become a space of reflection. 

Ultimately, AI makes of humanity what we choose to give it. My hope is to help shape AI as a poetic instrument of remembrance and possibility; one that invites humanity to see itself more clearly, and perhaps more responsibly.

Refik Anadol’s ‘Archive Dreaming’ installation will be open daily at the Schwarzman Centre from
25 April – 24 May 2026. 

Professor Edward Harcourt MBE 

Professor of Philosophy, Faculty of Philosophy, University of Oxford 

“ …whether human mindedness can be reproduced without the familiar materials of flesh and blood matters…” 

Much of the excitement, not to mention the anxiety, about AI has to do with the supposed similarities between AI and ourselves. What so often goes unnoticed is that the very term ‘artificial intelligence’ smuggles in a philosophical assumption that there’s a single characteristic that’s possessed by both humans and machines: intelligence. 

How much human territory could also be occupied by AI is a deep and urgent question. But another question in AI ethics that’s just as profound is the reverse: how the development of AI turns the spotlight on what is unique to us as humans. This was a worthwhile question before AI was even thought of — indeed, answering it could arguably be what humanism is all about. But progress in AI urges us to confront it like never before. 

For example, consider the everyday phenomenon of telling somebody something. If telling is about transferring information from one place to another, there is no apparent obstacle to machines telling us things because they are sources of information. 

But now consider the case of Jack and Jill. They are married, but Jack is having an affair. Jill knows this, and Jack knows Jill knows. In fact, Jill knows Jack very well so she knows that he knows she knows. So, there’s no relevant information Jill doesn’t already have. But all this could be true without Jack having told Jill anything. 

So maybe telling is not just about information transfer but about one person opening their mind to another, a transaction more like shaking hands or catching someone’s eye. That suggests it might be a humans-only affair. Similar questions arise about trust, friendship and a range of other transactions where ‘humans in the loop’ is valuable beyond the results it achieves. 

That’s not to say that all the most urgent ethical and practical questions raised by AI depend on fundamental questions about human mindedness. AI could shoulder us aside in some contexts even if it is fundamentally not like us: nobody needed to think the spinning jenny was ‘conscious’ to explain why it put a generation of yarn spinners out of work. Similarly interacting with a chatbot could cheer people up even if it lacks the value of a human-to-human conversation. 

But whether human mindedness can be reproduced without the familiar materials of flesh and blood matters: if it can’t, in areas like psychotherapy, or care for children and the elderly, or simple companionship, we may be irreplaceable. But we need better models of humanity to understand why.