AI & information integrity: Seeing is no longer believing | Sam Stockwell | TEDxLondonBusinessSchool

0:08 – 4:48

The Deepfake Quiz That Humbled the Room

Sam Stockwell opens by acknowledging the audience's AI fatigue, then throws down a challenge. Can you spot AI fakes? A few confident hands go up. He shows six images in rapid succession: mountaineers, a fluffy cat, a seagull, a fire hydrant. The audience whispers guesses to neighbors. Then comes the reveal. The first mountaineer photo? Real. The second? AI-generated. The cat? AI. The seagull? Also AI. The fire hydrant? AI again. He asks for a show of hands. Did anyone get full marks? Two hands creep up from a packed auditorium. The point lands hard. If trained eyes struggle today, what happens in two or three years when AI becomes photorealistic? Stockwell warns we may reach a point where discerning AI content with human eyes becomes virtually impossible. The stakes feel low with cats and fire hydrants, but they skyrocket when AI starts replicating people. As the gap between AI-generated and human-created content narrows, the line between fact and fiction blurs. And anyone can be deepfaked now, not just celebrities or politicians. Stockwell himself appears on screen as a deepfake, though he notes this is a rare case where he actually consented.