Human Intelligence in the Age of AI

When we hear about new developments in the world of artificial intelligence, the big bangs are always about novel capability breakthroughs, often associated with unthinkable amounts of new training data or massive additions of computing power. Then, invariably, the doomsayers follow, sounding the warning bells about AI as a harbinger for the end days of mankind. While debating what AI can do versus how it will destroy us, we tend to skip over the very thing we are so terrified of losing to machines – human intelligence.

Human intelligence and exactly how it fits and adapts alongside a fast-developing artificial intelligence, was at the core of the conversation at the AI in Education conference at King’s College London on Monday, 3 March, 2025. From an education perspective, this was refreshing, because considering the human alongside AI seemed to re-orient all the attendees to the very human goal of educating our children in the best way possible.

Professor Sylvie Delacroix, Director of the Centre for Data Futures at Kings College London, had the most poignant example of how human intelligence could be harnessed to interact with a hypothetical model of AI. She envisaged a physics student, slightly discouraged by the fact that she was given homework of lesser difficulty than her peer, engaging with the system. First, the student asked why. The AI reasoned that she had performed worse on her exams. Instead of fully accepting the answer, the student continued to ask what the harder problems looked like. Finding them challenging but not completely out of her comfort zone, the student asked the system to put her on the more difficult pathway, and it obliged.

This type of interaction retains student agency. It also caters to the strengths of a well-programmed AI tool with multiple learning pathways for the same subject. Overall, it presented a fantastic example of the way that the two intelligences can work together for educational betterment.

Of course, in that example, the student needs to be aware of the AI systems with which they would be working, and also confident enough in subject matter knowledge to help themselves learn effectively. This brings us to another topic which Professor Rose Luckin of University College London touched upon.

Professor Luckin spoke about how to deepen human intelligence in the age of AI. She posited that AI can enhance human intelligence, but only if we use it conscientiously. As long as we maintain an understanding of learning based in cognitive science to further education, AI won’t pose a problem for education. But if we attempt to replace human intelligence with artificial intelligence, we will be doing ourselves a disservice, Professor Luckin said.

Professor Neil Lawrence from the University of Cambridge took the idea of human and AI interaction to a different level, arguing that machine processing greatly surpasses our brain’s capabilities in terms of speed. But as machines strip away what were previously human tasks, what’s left is the atomic human – quintessential humanity. Professor Lawrence continued to say that what makes us truly human may in fact be our flaws.

Though every presentation and conversation was thought-provoking, two larger themes emerged from the conference. The first was that knowledge is, and will remain, important to learn in schools. The education system may adopt different methods of assessment and curricula may adapt to new technologies in the world like AI, but the ability to use AI to access knowledge will never be the same as knowing something, just as the ability to Google something does not mean the search engine user immediately gains that information.

The second theme was that a large portion of the ‘education’ provided in schools is social, and that cannot be replaced by AI. The danger is that AI might erode some of the social learning aspects of schools that bind society together. Even with something as innocuous and positive-sounding as ‘personalised learning pathways,’ there is a slight danger of altering or removing the type of collective knowledge transmitted by teachers and among students in a classroom that is so crucial for the development of pupils.

That is human intelligence. And it is the mandate of schools to sharpen, develop, and provide the fertile ground for it to flourish, with or without AI.

Luke Kemper is Insight and Intelligence Lead at HEP. He graduated from the University of Cambridge with an MPhil in Education, Globalisation and International Development. Before that, he worked for seven years as a university lecturer and high school teacher in China and Poland.