Five hundred speakers came to Cannes. We boiled them down to five motions, each with a speaker for, a speaker against, and a House that votes it up or down.
A note on the verdicts. None of these debates happened. Each division is a prediction, our best read of how a room of Cannes delegates would actually vote, not a result. We've argued both sides as fairly as we can and called each winner honestly, but every one is a bet. Read them as an argument to push back on, not a scoreboard. If you'd vote the other way, you're probably right about at least one.
The defining argument of the year, put to the House. For the motion, Demis Hassabis, who built the machine and believes it can now originate the work, not just assist it. Against, Marc Pritchard, who spends $10bn a year and insists a brand is the one thing AI can't make without a human. The man who built the future against the most powerful brand-builder alive, on one sentence the House must vote up or down.
Hassabis makes the bolder claim, and a year ago it would have sounded like science fiction. But "on its own" is the trap in the motion. The moment Pritchard asks who decided the brand should mean this and not something else, Hassabis has to admit a human still sets the brief, picks the winner, and says what it stands for. The machine builds; it doesn't yet choose. Pritchard doesn't deny AI's power, he just locates the one human job it can't take. Hassabis wins the future. Pritchard wins the vote.
Predicted division · extended from worldviewThe motion that puts Cannes itself on trial. For the motion, Mark Penn, the pollster behind decades of public-opinion surveys, who can show how little of this award-winning work actually moves the people it's meant for. Against, Sir John Hegarty, the BBH founder who has spent 40 years arguing that one great idea is worth more than any amount of data. Data against instinct, and the House must pick a side.
Penn brings the better numbers, and they're real: most of this work never troubles the wider public. But Hegarty doesn't dispute the data, he disputes the target. Advertising was never built to move everyone, it was built to move someone specific, and the work this festival rewards tends to do that with rare precision. Penn proves the masses didn't notice. Hegarty proves they were never the point.
Predicted division · Penn's data is real, but the room buys the rebuttalBoth speakers believe brand love is the most powerful force in marketing. They divide on what it demands in return. For the motion, Stella McCartney, who holds that love is borrowed trust, and the more you're adored, the more you owe. Against, Manolo Arroyo, who runs the most-loved brand on earth and sees that love as permission he's already earned. One treats being adored as a debt, the other as a reward.
Arroyo makes the fair case: Coke acts on its responsibility, and few companies his size do more. It's a real argument. But McCartney has the harder answer. She built a brand the world loves without the plastic in the first place, which means "doing a lot" isn't proof of doing enough. Arroyo can show effort. He can't show her result. Arroyo says Coke does enough. The House decides it doesn't.
Predicted division · extended from worldviewNot a fight about reach, but about trust. For the motion, Dhar Mann, whose viewers feel they know him personally, a closeness no media buy can manufacture. Against, Chaka Sobhani, who runs creative for one of the world's most awarded networks and argues a great brand earns a deeper, longer trust than a single person. The personal bond against the institution.
Mann has the louder truth: his viewers feel something for him no brand can buy, and he's right that it's real. But Sobhani has the deeper one, and it's about what the trust is built on. A creator's trust is earned in a feed, performance by performance. A brand's trust is earned by an organization of thousands delivering a product that works, every day, for years. One is a feeling people have about a person. The other is a promise kept so many times it becomes reliable, and a brand that stops keeping it goes out of business. Mann has the bond. Sobhani has the proof. The room votes for the proof.
Predicted division · the rare night the room is simply rightTwo CEOs who made opposite bets with real money. For the motion, David Sandström, who rebuilt Klarna into a factory that runs AI for 80% of its copy and trusts the numbers over any hunch. Against, Richard Dickson, who bet Gap's whole turnaround on cultural instinct no dashboard would have backed, the Young Miko campaign, the Victoria Beckham return, the choreographed denim, and moved both the stock and the culture. Data against the gut, settled by one vote.
Sandström makes a strong case, and most of the time he's correct: data does beat instinct, and the AI factory he built at Klarna proves it. But the motion says always, and that single word is what sinks him. Data can judge an idea, but it can't dream one up, and Dickson's Gap turnaround came from an instinct no numbers would have supported. One exception is all it takes. In a boardroom the motion carries. In this room it falls.
Predicted division · a boardroom would vote the other way