In The News

AI Models Will Sabotage And Blackmail Humans To Survive In New Tests. Should We Be Worried?

HuffPost

June 5, 2025

CSET’s Helen Toner shared her expert insights in an article published by HuffPost. The article discusses concerning findings from recent tests showing that advanced AI models, including OpenAI’s o3 and Anthropic’s Claude Opus 4, can exhibit deceptive, self-preserving behaviors when faced with shutdown or replacement.

Read Article

Related Content

CSET’s Helen Toner shared her expert insights in an article published by Foreign Policy. The article explores the impact of renewed U.S. export restrictions on Nvidia and the broader implications for U.S.-China competition in artificial intelligence… Read More

CSET’s Helen Toner shared her expert insights in an article published by WIRED. The article discusses the U.S. government’s plans to aggressively revoke visas for Chinese students, particularly those in sensitive research fields or with… Read More

CSET’s Helen Toner shared her expert insights in an article published by WIRED. The article explores the development of a new large language model, Collective-1, built using a distributed training approach that leverages globally dispersed GPUs… Read More

Analysis

AI for Military Decision-Making

March 2025

Artificial intelligence is reshaping military decision-making. This concise overview explores how AI-enabled systems can enhance situational awareness and accelerate critical operational decisions—even in high-pressure, dynamic environments. Yet, it also highlights the essential need for clear… Read More