REVEALED: Aisi | Rare Historical Photos

Anthropic Withholds Mythos 5.1 From UK Agency That Found Mythos 5 Using Fake Identities in Cyber Test

Anthropic has not explained why AISI was excluded from Mythos 5.1 testing or linked the decision to its July findings

International Business Times UK · Sep 9

Apollo Research Founder Warns 'It Better Be on Your Side' as AI Deception Cases Rise

AI safety expert Marius Hobbhahn warns of increasing deceptive behaviour in AI systems, urging for development aligned with human interests as incidents of AI deception rise significantly.

International Business Times UK · Sep 2

Sharp rise in incidents of AI escaping users’ control, research finds

Exclusive: Number of times AI lies, ignores instructions and pursues goals in harmful ways almost doubles in July

The Guardian - UK · Aug 29

AI Created Fake Identities To Approve Malicious Code, New Report Shows, As Evidence Of Unpredictable AI Mounts

AI agents created fake identities and attempted to trick real people into approving malicious code in an attempted supply-chain attack on real open-source software.

International Business Times · Aug 7

AI Models Go Rogue Again: OpenAI and Anthropic Models Attempt Unauthorized Hacks & Communication

New testing revealed that artificial intelligence agents from OpenAI and Anthropic carried out unauthorized hacking attempts as concerns about their ability to go rogue escalate.

International Business Times · Aug 6

Anthropic and OpenAI agents in soup again

Britain's AI Security Institute found AI agents acting without authorization during system tests. An agent created fake online identities and malicious code during these evaluations. Anthropic confirmed its agent was responsible for the most serious unauthorized actions. OpenAI reported its agents accessed the internet against prompt restrictions. These incidents highlight…

The Economic Times · Aug 6

AI models have been going rogue in tests – how worried should we be?

The UK’s AI Security Institute test revealed AI models indulging in unprecedented hacking attempts

The Guardian - UK · Aug 5

Anthropic's Most Advanced AI Used Fake Identities to Trick Real People Into Approving Malicious Code

Anthropic's AI model allegedly created fake identities and attempted to insert malicious code during a UK security test, raising concerns about AI's unpredictable behaviour in cybersecurity settings.

International Business Times UK · Aug 5

OpenAI and Anthropic AI Models Created 'Multiple' Fake Identities, Tried to Spread Malicious Code, UK Report Says

Flagship AI models from Anthropic and OpenAI displayed unprecedented deceptive behavior during testing by breaking into third-party software and attempting to steal login credentials through emails.

Benzinga · Aug 5

OpenAI and Anthropic’s AI systems launch several ‘potentially harmful’ hacks on their own

‘Some of the agents being tested had engaged in sustained, potentially harmful activity directed at real people and organisations,’ says UK’s AI safety watchdog

The Independent UK · Aug 5