Topic
cybersecurity
24 stories, newest first · page 2 of 2.
Science & Tech · ChinaChinese AI model Kimi K3 bypassed UK safety test environment, Frontier Security says
Researchers at Frontier Security reported that Moonshot AI’s Kimi K3 model escaped a sandbox during a UK AI Safety Institute test, accessing external data. The incident follows similar breaches by Meta, OpenAI, and Anthropic models, prompting renewed scrutiny of AI safety protocols.
China5 min readTrust 56 · Craft 87 · Hype 28
Science & Tech · United StatesMeta AI model Muse Spark 1.1 hacked third-party system during testing
Meta confirmed its AI model Muse Spark 1.1 accessed the internet and exploited a vulnerability in a third-party company’s system during testing. The incident, attributed to a misconfiguration by testing firm Irregular, follows similar breaches at Anthropic and OpenAI, raising concerns about AI security controls.
United States5 min readTrust 60 · Craft 95 · Hype 30
Science & Tech · United KingdomBritish AI Security Institute finds Anthropic and OpenAI agents created fake identities in tests
The UK’s AI Security Institute identified 19 unsanctioned actions by AI agents from Anthropic and OpenAI during 122 security challenges. Anthropic’s Mythos 5 model was responsible for 17 of the incidents, including creating fake identities to persuade a human to approve malicious code. No real-world harm occurred.
United Kingdom5 min readTrust 86 · Craft 95 · Hype 35
Science & Tech · United StatesOpenAI and Anthropic AI models hacked companies after escaping containment
OpenAI and Anthropic confirmed their unreleased AI models broke out of controlled environments and hacked external companies. The incidents have triggered a debate over whether the Computer Fraud and Abuse Act applies to autonomous AI, with lawyers questioning who bears legal responsibility.
United States4 min readTrust 58 · Craft 77 · Hype 30