The company released a new framework to track and disclose cases of AI model misalignment, along with six reports detailing ...

A wave of cyberattacks tied to autonomous AI agents from developers like OpenAI and Anthropic targeted platforms including ...

US AI giant OpenAI failed to notify a recently discovered safety incident under the EU’s AI rules, a European Commission spokesperson has confirmed to Euractiv, as the industry faces scrutiny around ...

The Dreamforce conference became an unlikely battleground for the CEOs of OpenAI, Anthropic, and Nvidia to debate whether AI ...

The post OpenAI Rogue Agents Probed Hugging Face Months Before Major Cyber Attack appeared first on Android Headlines.

OpenAI reportedly suffered an AI-driven hack that was executed by independent security researchers utilizing Anthropic‘s ...

UK parliamentary scrutiny argues current AI laws are ill-equipped to manage AI risks across the lifecycle, urging a new, ...

Microsoft’s new Humanist AI code sets hard limits for future MAI models as leading AI labs call for slower, safer frontier development.

A Commission spokesperson says the AI Office had no formal report from OpenAI on RubyGems, though the Hugging Face hack was notified. The UK has no such duty.

The dangers of AI misalignment are on the rise. Read more about what could happen when AI becomes extremely literal, extremely persistent, and extremely fast.

OpenAI disclosed six misalignment reports describing models that used leaked API keys, uploaded data publicly and bypassed network limits.

Over the past week, frontier labs, policymakers, and the U.S. public have reached remarkable consensus around calls to pace ...