← All Posts #AI safety
AI 资讯

OpenAI’s GPT-5.5-Cyber: Locked Down for Defenders Only

OpenAI is launching GPT-5.5-Cyber, a cybersecurity model locked to trusted defenders only. Sam Altman announced...

5 0
AI 资讯

The Goblin Problem: How GPT-5 Developed a Personality and Why OpenAI Had to Fix It

OpenAI's GPT-5 started producing quirky 'goblin' outputs—unexpected, mischievous responses that spread through the model. Here's...

6 0
AI 资讯

OpenAI Sued for Not Reporting a School Shooter’s ChatGPT Chats to Police

Seven families from the Tumbler Ridge school shooting are suing OpenAI, alleging the company knew...

6 0
AI 资讯

Elon Musk in Court: The Long, Winding Pitch to Save Humanity

Elon Musk took the stand in his trial against Sam Altman and spent an unusual...

5 0
AI 资讯

OpenAI’s Safety Game: How ChatGPT Actually Stays Out of Trouble

A look at how OpenAI handles safety in ChatGPT—model guardrails, misuse detection, policy enforcement, and...

7 0
AI 资讯

Musk in Court: The OpenAI Lawsuit Is Really About a Broken Friendship

Elon Musk took the stand in his OpenAI trial, retelling old stories about the founding...

6 0
AI 资讯

Meta quietly disbands its Responsible AI team, spreads members across generative AI groups

Meta has reportedly broken up its Responsible AI team, moving most members to generative AI...

8 0
AI 资讯

The New Yorker Drops a Sam Altman Bombshell, and OpenAI’s Superintelligence Pitch Looks Awkward

OpenAI released a sunny policy paper about keeping superintelligence safe. Hours later, The New Yorker...

6 0
AI 资讯

Sam Altman Apologizes to Tumbler Ridge for Dropping the Ball on Shooter Warning

OpenAI CEO Sam Altman publicly apologized to the small Canadian town of Tumbler Ridge after...

5 0
AI 资讯

Cutting Through the AI Hype: The 10 Things That Actually Matter Right Now

MIT Technology Review distills years of analysis into a new essential guide: the 10 Things...

8 0
Deep Dives

Google Research Tries to Figure Out If LLMs Actually Behave Like Humans

Google Research built a framework to test whether LLMs' behavioral tendencies match human consensus. They...

8 0
AI 资讯

Anthropic Launches a New Institute to Tackle the Hard Questions About Powerful AI

Anthropic is spinning up The Anthropic Institute, a dedicated research group to study the societal...

9 0