Sal Khan built Khan Academy from a handful of YouTube tutorials into one of the world's largest free education platforms, and more recently built Khanmigo, an AI tutor. At UNESCO's Digital Learning Week 2026, he sat down for a fireside chat with UCL's Wayne Holmes — an academic whose entire field is scrutinising AI's ethics in education. That pairing, on that stage, is worth understanding even without a transcript.

We want to be honest about what we know here: UNESCO's programme confirms this conversation happened. We don't have a recording, transcript, or direct quotes from either Sal Khan or Wayne Holmes, and we're not going to invent any. What we can do honestly is explain who these two people are, why the pairing itself is meaningful, and what questions a conversation like this is built to raise — questions worth asking of any AI tutoring product, including the ones you're choosing between for your own child.

Who was on stage

Sal Khan founded Khan Academy in 2008 and has since built Khanmigo, an AI tutor aimed at guiding students through problems rather than just answering questions. He's arguably the most recognisable figure in AI-assisted education globally — someone with a genuine track record of free, accessible teaching, and now a direct commercial stake in an AI tutoring product.

Wayne Holmes is a UCL academic whose work centres on the ethics of AI in education — the person whose job, structurally, is to ask hard questions of exactly the kind of product Sal Khan builds. A "fireside chat" pairing the two isn't a keynote where one person talks uninterrupted; it's a format built for scrutiny and genuine back-and-forth.

Why the pairing itself is the story

Here's what's actually significant, independent of anything either of them said: an AI-tutoring industry figure agreed to be questioned, on a UN body's own stage, by an academic whose field exists to be sceptical of exactly his kind of product. That's a meaningfully different posture from marketing copy on a website, or a keynote with no room for pushback.

It also mirrors something worth naming honestly: every company that builds an AI tutor — including aitutors.me — has a commercial interest in that tutor being used. That doesn't make the tool bad, but it does mean the ethical scrutiny has to come from somewhere other than the company's own marketing. A fireside chat with an AI-ethics academic, at a global education body's flagship event, is one visible form that scrutiny can take.

The questions a conversation like this has to answer

Without a transcript, we can still name the questions this pairing is structurally built to raise — because they're the same questions any serious AI-ethics academic asks of any AI tutoring product:

1. Does it protect the moment of thinking, or hand over the answer? This is the single most important design question in AI tutoring. A tool that gives a struggling student the final answer might feel helpful in the moment and teach nothing by the end of term. We've written about this directly in why tutors don't give answers.

2. What happens to a child's data? Any AI product that a minor interacts with regularly is collecting something — conversation logs, performance data, sometimes more. A UK parent has every right to ask exactly what's stored, for how long, and who can see it.

3. Who actually gets access? Free, high-quality AI tutoring changes outcomes most for students who can't otherwise afford a human tutor — which is exactly why equitable access is a live ethical question, not just a nice-to-have.

What this means for you

You don't need to have watched the Khan–Holmes conversation to use it well. The lesson is simpler than that: any AI tutoring product you consider for your child should be able to answer those same three questions plainly, in writing, without dodging. If a product can't tell you clearly how it handles the thinking-versus-answering line, what happens to your child's data, or who it's actually reaching, that's worth noticing.

For where aitutors.me stands on these same questions, see our piece on is AI tutoring safe for kids in the UK, and for the wider UNESCO context, our Digital Learning Week 2026 overview or the complete FAQ.

FAQ

What did Sal Khan discuss at UNESCO Digital Learning Week 2026?

Sal Khan, founder of Khan Academy and its AI tutor Khanmigo, held a fireside chat with UCL academic Wayne Holmes. UNESCO's programme confirms the pairing took place; we don't have a transcript or direct quotes, so we can only speak to the kind of questions such a conversation would plausibly cover, not what was actually said.

Who is Wayne Holmes?

Wayne Holmes is a UCL academic known for his work on AI ethics in education — the kind of person whose job is to ask hard questions of AI-in-education products, including ones built by people sitting next to him.

Why does it matter that an AI-tutoring founder sat down with an AI ethics academic?

Because it signals the industry answering to scrutiny in public, at a UN body's own event, rather than only marketing to parents directly. That's a healthier dynamic than an unchallenged sales pitch — whichever specific product is being discussed.

Duke Harewood built aitutors.me for his own KS3 daughter, and holds the same three questions this article raises against our own product every time we change it. Published 11 September 2026.