Name: What if AI could spot your lies?
Uploaded: 2025-01-27T15:50:10Z
Duration: 15 min
Description: Humans are terrible at detecting lies, says psychologist Riccardo Loconte ... but what if we had an AI-powered tool to help? He introduces his team’s work successfully training an AI to recognize falsehoods in certain contexts, laying the groundwork for a world where everything from national security to social media is a little bit safer — and a bit more ethically complicated.

Riccardo Loconte: What if AI could spot your lies?

What if AI could spot your lies?

Humans are terrible at detecting lies, says psychologist Riccardo Loconte ... but what if we had an AI-powered tool to help? He introduces his team’s work successfully training an AI to recognize falsehoods in certain contexts, laying the groundwork for a world where everything from national security to social media is a little bit safer — and a bit more ethically complicated.

Most people think they can spot deception, yet studies show humans are only as accurate as flipping a coin. Researchers trained AI language models to identify lies and achieved up to 80 percent accuracy, far surpassing human performance. While this technology could revolutionize fields from law enforcement to social media, it also raises concerns about blind trust in algorithms and the erosion of interpersonal trust. The future of lie detection may be powerful, but it must remain interpretable and accountable.