Oasis Securityは、OpenClawのゲートウェイに任意のWebサイトからローカル接続を悪用しAIエージェントを乗っ取れる欠陥を発見したと公表した。開発側は高深刻度と判断し24時間以内に修正版を公開した。

AnthropicとMozillaの共同調査で、AIモデル「Claude Opus 4.6」がFirefoxから22件の脆弱性を見つけた。厳しく検証されてきたソフトでもAIが新たな弱点を洗い出せる可能性を示した一方、攻撃より防御での実用性が先行している点も注目される。

AnthropicのFrontier Red TeamとMozillaがAIを用いた脆弱(ぜいじゃく)性検出に関する連携を行い、Claude Opus 4.6がわずか2週間の調査でFirefoxについて計112件の報告を提出し、その中から22件の脆弱性が確認されたことを報告しました。この成果は、AIが大規模なコードベースの安全性を極めて高い速度で検証し、強化できる可能性を実証するものです。

This high-severity Chrome Gemini vulnerability lets malicious extensions spy on your PC ...

In a speech to Congress, Argentinian President Javier Milei hs said that he wants Argentina's alliance with the US to become ...

Canadian experts need to assess ChatGPT conversations that have been flagged for signs that users intend to cause imminent harm to determine whether to alert law enforcement, federal AI Minister Evan ...

It's a lifesaver.

The now-patched flaw is the latest in a growing string of security issues with the viral AI tool, which has seen rapid adoption among developers.

The iPhone crypto exploit Coruna targets outdated iOS devices, stealing wallet seed phrases through compromised sites, ...

Typical of an unconventional presidency, the Trump administration waited more than 48 hours to communicate to the American ...

Millions installed 'productivity' Chrome extensions that became malware after acquisition. Here's how browser extensions became enterprise security's weakest link.

Sam Altman, the chief executive officer of OpenAI, one of the most valuable privately held companies in the world, will delivery an apology to families in Tumbler Ridge after hearing about the impact ...