The Future with Hannah Fry: Rethinking Intelligence in the Age of AI
Written by Rebecca Uffindell Wed 4 Mar 2026
Professor Hannah Fry brought a lively and thought-provoking session to the Tech Show London Mainstage, using current scientific research to explore what artificial intelligence is revealing about identity, communication and consciousness.
Rather than speculate about distant futures, Fry focused on developments already reshaping how intelligence is understood. Her argument was simple but unsettling: many of the traits long considered uniquely human may not be as exclusive as once assumed.
Pigeons, Pathology and Collective Intelligence
Fry opened with an experiment in which pigeons were trained to diagnose breast cancer from pathology slides. Individually, the birds achieved 86% accuracy. When their responses were averaged collectively, that figure rose to 99% – matching trained pathologists.
The example challenged comfortable assumptions about expertise.
“Maybe artificial intelligence is suddenly doing so well not because they’re becoming more human,” she said, “but maybe it’s because the things that we thought made us special actually turned out to be pigeon work.”
Her point extended beyond the novelty of the experiment. Tasks often framed as hallmarks of human reasoning may be closer to pattern recognition than we would like to admit. When viewed this way, the rapid progress of AI systems appears less mysterious and more inevitable.
What Makes Us Human?
Fry then turned the lens toward the audience.
In a live Turing-style exercise, participants submitted single words intended to signal humanity. The activity drew on research analysing thousands of similar word submissions.
Many chose emotionally charged words such as “love”. Yet the most reliable word for distinguishing humans from machines turned out to be far less poetic: “poop.”
“In four letters, what this word has managed to do is to find that dividing line,” Fry said.
Her conclusion was deliberately provocative. Biological experience – embodied, physical and occasionally inelegant – may define human distinctiveness more than abstract reasoning ever did.
She described this insight as “very deeply profound,” arguing that discussions about AI often overlook the role of physical experience in shaping consciousness.
The Consciousness Debate
Fry outlined two prevailing theories about artificial consciousness. One suggests that consciousness emerges naturally once systems reach sufficient intelligence. The other proposes that consciousness arose through evolutionary pressures demanding social awareness and internal state modelling.
“You will not see consciousness emerge in artificial systems unless you put them through a series of evolution,” she argued.
Her position places consciousness within a biological and social context rather than a purely computational one. As AI systems grow more capable, this distinction will shape future debates about rights, agency and responsibility.
The question is no longer abstract. Advances in generative models and autonomous systems are pushing philosophical discussions into practical territory.
Speaking to Animals
One of the most ambitious parts of the session examined AI’s role in decoding animal communication.
Fry described how modern language models embed words in multidimensional space, where meaning emerges from relationships between terms. These techniques underpin contemporary translation systems.
Applying similar methods to animal vocalisations could create what she described as an “AI Rosetta Stone.”
Examples included:
Dogs using button-based systems to communicate needs and internal states.
Research showing elephants respond to recorded “bee” alarm calls and may use individual names.
Efforts to analyse decades of animal vocalisation recordings with machine learning tools.
“We need to learn how to speak dog,” Fry said.
The implication is expansive. Translation across species would alter how intelligence is defined, how empathy is extended and how non-human cognition is valued.
Rethinking Intelligence
Throughout the session, Fry returned to a broader theme: intelligence exists in multiple forms — biological, collective and artificial. The rise of AI is forcing a reassessment of assumptions long treated as fixed.
Humans often rely on a narrow definition of intelligence rooted in language and abstraction. Machine learning systems and biological research are steadily expanding within that frame.
As AI advances, boundaries between human and non-human intelligence are becoming harder to defend through intuition alone. Distinctions that once felt obvious now require explanation.
The conversation did not hinge on prediction. It centred on perspective. Artificial intelligence is not only accelerating capability; it is exposing how incomplete our understanding of intelligence has always been.
The future may not depend on machines becoming more human. It may depend on humans reconsidering what intelligence truly encompasses.
Written by Rebecca Uffindell Wed 4 Mar 2026