Your chatbot says you're brilliant – that's a problem

on
Plus: Why richest man in early 1800s Philly had his socks mended ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌
The Conversation

Welcome to the Saturday edition of The Conversation U.S.’s Daily newsletter.

A little flattery goes a long way – even if it’s coming from AI. Chatbots, like people, sometimes tell us what we want to hear: that a question is deep, or an idea is clever.

Nir Eisikovits, director of UMass Boston’s Applied Ethics Center, and Cody Turner, a philosopher at Bentley University, explain how the way AI is developed can lead it to prioritize the user’s approval over accuracy. That may not seem too consequential for the types of little questions we often throw at chatbots.

But as the stakes get bigger, it has serious consequences for truth and trust. The risks of AI sycophancy aren’t just about being able to tell fact from fiction and making decisions accordingly. Eisikovits and Turner argue it’s also dangerous for psychology and for democracy – not to mention our ability to navigate “the world of human relationships, where friction, disagreement, boredom and different opinions than your own are prevalent.”

This week we also liked stories about the Supreme Court case about law enforcement tracking crime suspects via other people’s phone locations, the state of synthetic biology science after the death of genomics pioneer J. Craig Venter, and the dangers of bringing ‘medical freedom’ to the U.S. military’s vaccine policy.

Did somebody forward this email to you? Subscribe to our daily and weekly newsletters here.

Molly Jackson

Religion and Ethics Editor
Sycophancy eats away at truth and trust. Andriy Onufriyenko/Moment via Getty Images

AI chatbots can prioritize flattery over facts – and that carries serious risks

Nir Eisikovits, UMass Boston; Cody Turner, Bentley University

Everyone likes being told they’re clever, even if it’s coming from an AI chatbot. But their sycophancy has serious consequences for truth and trust.
Do these look like the socks of the wealthiest man in America? Courtesy of Girard College History Collections

Stockings once worn by Philly’s wealthiest man show the value of women’s mending in early America

Emily J. Whitted, UMass Amherst

Stephen Girard’s silk stockings were repaired by multiple women in his household who made his enormous wealth possible.
A TV displays U.S. President Donald Trump’s prime-time address on the war in Iran inside a Cheesecake Factory on April 1, 2026, in Washington, D.C. Anna Moneymaker/Getty Images

Why the 60-day War Powers Resolution deadline doesn’t actually constrain presidents

Jasmine Farrier, University of Louisville

Donald Trump’s unilateral move to engage in military action against Iran isn’t unprecedented; Presidents Obama and Clinton directed U.S. military engagements without explicit congressional approval

Synthetic biology promised to rewrite life – with the death of its pioneer, J. Craig Venter, how close are scientists?

André O. Hudson, Rochester Institute of Technology

Advances in genetic engineering have enabled researchers to seek ways to program new life. But has synthetic biology actually changed medicine and the environment, nearly two decades on?

‘A study showed…’ isn’t enough – scientific knowledge builds incrementally as researchers investigate and revisit questions

Jeffrey A. Lee, Texas Tech University

Each study adds a piece to the puzzle of scientific knowledge. But any one study on its own doesn’t tell you all that much.

Your local storm forecast is likely based on weather miles away – we’re trying to bring it closer to home

Chris Vagasky, University of Wisconsin-Madison

Don’t blame the weatherman. A meteorologist explains where your local weather forecasts really come from, and how storm scientists are trying to fill the gap.

Sora’s downfall signals broader problems with AI’s creative utility

Ahmed Elgammal, Rutgers University

AI’s apparent inability to generate interesting and unique images and videos is becoming harder to ignore.

The Conversation News Quiz 🧠