- Snopes
- Posts
- AI Models, Children's Cancer Research & the FDIC
AI Models, Children's Cancer Research & the FDIC
Plus: 'Jingle Bells' Wasn't Written as a Christmas Song. Here's the Real History
TUESDAY EDITION
Snopes Debunker
AI Models Were Caught Lying to Researchers in Tests — But It's Not Time to Worry Just Yet
In early December, claims circulated online that researchers had caught artificial intelligence models — specifically OpenAI's o1, which is used in ChatGPT Pro, a popular AI chatbot — lying in tests to avoid being shut down. The claims persisted later into the month, with users specifically noting the results in relation to ChatGPT.
The claims stem from research carried out by Apollo Research, an AI safety organization. The organization found five out of the six tested AI models "showed scheming capabilities" in tasks set by researchers that "incentivize scheming." OpenAI's o1 model was capable of carrying out all the scheming behaviors tested, according to the research. Snopes reporter Laerke Christensen investigated.
Critics Say Musk's Influence on Trump Led GOP to Cut Children's Cancer Research. Here's What We Know
A U.S. government shutdown loomed after a Trump-backed spending bill failed to pass in the House.
Yes, Biden Commuted Sentences of Chinese Spies and Sex Offender. Here's What Else We Found
The commutations were separate from the more than 1,000 people whose sentences the U.S. president commuted on Dec. 12.
Project 2025 Proposes Eliminating the FDIC? Not Exactly
A 2017 report by conservative think tank The Heritage Foundation suggested "government-provided deposit insurance should be phased out fully."
Unpacking Claims That High Radiation Levels in New York Are Connected to Drone Sightings
Reports of high radiation levels in New York City were not corroborated by official figures.
Staff Pick
'Jingle Bells' Wasn't Written as a Christmas Song. Here's the Real History
Popular myths about the classic holiday tune have circulated for decades, obscuring its origins on the minstrel stage.