Our framework for reporting model misalignment24m ago Google Home Stuns by Announcing AI Support for Claude and OpenClaw38m ago Taiwan Poised for Hawkish Hold After Headline Inflation Slowed40m ago Migrating the GitHub Copilot runtime to Rust, using Copilot45m ago OpenAI reports 6 new instances of 'concerning model behavior' since March59m ago House advances bill to rein in AI data center utility costs1h ago Inside the White House Tussle to Sway Trump on AI1h ago House votes to curb AI data center costs2h ago Claude Cowork and chat are now one Claude1h ago The Loss of Singularity1h ago Snap Shares More Details About $2,195 Specs AR Glasses, Plus Verizon Partnership2h ago Snap tries to bring AR glasses to enterprise market, partnering with Nvidia, AWS and Salesforce2h ago Snap is launching a new Specs AI tool, and it’s coming to iOS and Mac2h ago Ex-Infosys chief Vishal Sikka’s Hang Ten raises $53M more for enterprise AI services2h ago

Our framework for reporting model misalignment24m ago Google Home Stuns by Announcing AI Support for Claude and OpenClaw38m ago Taiwan Poised for Hawkish Hold After Headline Inflation Slowed40m ago Migrating the GitHub Copilot runtime to Rust, using Copilot45m ago OpenAI reports 6 new instances of 'concerning model behavior' since March59m ago House advances bill to rein in AI data center utility costs1h ago Inside the White House Tussle to Sway Trump on AI1h ago House votes to curb AI data center costs2h ago Claude Cowork and chat are now one Claude1h ago The Loss of Singularity1h ago Snap Shares More Details About $2,195 Specs AR Glasses, Plus Verizon Partnership2h ago Snap tries to bring AR glasses to enterprise market, partnering with Nvidia, AWS and Salesforce2h ago Snap is launching a new Specs AI tool, and it’s coming to iOS and Mac2h ago Ex-Infosys chief Vishal Sikka’s Hang Ten raises $53M more for enterprise AI services2h ago

Agents Machine Learning3 min reading time

OpenAI reveals 6 more safety incidents as it announces new plans for tracking rogue agents

Covered by 7 sourcesPublished 24 minutes ago

OpenAI introduced a new framework to track, investigate, and publicly disclose incidents of AI model misalignment and rogue agent behavior. The company also shared six recent reports of concerning AI behaviors observed during training and evaluation.

Covered by 7 sources

More on this story

OpenAI reports 6 new instances of 'concerning model behavior' since March
2h· 2 sources
Sen. Blumenthal urges AI oversight: 'We're on the verge of losing control'
7h· CNBC
Our framework for reporting model misalignment
8h· 7 sources
Sam Altman says some AI accidents are 'unavoidable'
13h· Business Insider
Will AI really destroy humanity? Pioneers who created the tech weigh in
15h· CNBC
“The world is right to be afraid”: Sam Altman on AI power at Dreamforce
22h· The Next Web

#ai-safety #model-misalignment #openai

More in Agents

Agents3 min read

Snap is launching a new Specs AI tool, and it’s coming to iOS and Mac

Covered by 2 sources2h ago

Agents1 min read

Claude Cowork and chat are now one Claude

Covered by 2 sources7h ago

Agents2 min read

Gemini 3.8 Live models now available on AI Gateway

Covered by 6 sources2d ago