Gradual Disempowerment from AI in Competitive Debating

LessWrong notes that competitive debating has long been a proving ground for sharp thinkers, from senators to AI researchers. While a 2019 face‑off saw IBM’s Project Debater lose to human champion Harish Natarajan, a 2026 study by Hackenburg and colleagues found modern AIs to be far more persuasive than a group of fifty‑six elite debaters in text‑based UK political conversations. The researchers also reported that when the AI’s output was limited to human‑level message length and speed, its performance was comparable to the top humans. Per the essay, this shift could hollow out debate’s apprenticeship role and alter its place as a feeder into frontier AI research and governance.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton