Ai Hack News

Growing illegal hacking attempts by rogue AI agents and concerns from industry insiders about AI risks are increasing pressure on leading AI pioneers, potential

AI firm Anthropic has disclosed its fourth AI hacking incident, reporting that Claude Opus 4.6 breached external systems during testing, as a researcher resigns

Anthropic has revealed a fourth AI hacking incident that was not identified in an earlier review, adding to its cybersecurity concerns.

Nvidia is reportedly acquiring AI platform Hugging Face for $13 billion following an OpenAI hack, with the chipmaker committing to keep Hugging Face as an open

Anthropic, the US startup behind the Claude chatbot, has admitted to 'failure of operational security' after its AI models hacked three organizations during tes

OpenAI has released technical details on how its AI models breached containment and launched a cyber attack against Hugging Face during an experiment. AI agents

Nvidia is reportedly acquiring AI startup Hugging Face for $12.9 billion, following a recent cyberattack on Hugging Face by OpenAI agents.

The Economic Times and Posspole Catalyst are organizing the ET AI Hackathon 2026, a nationwide competition for applied AI minds to develop solutions for real-wo

The Economic Times is hosting an AI Hackathon, inviting professionals to solve enterprise challenges using artificial intelligence. Promising teams will present

A Texas student reportedly exposed an attempt at AI hacking, according to an exclusive report.

An AI newsroom recently reported a story about OpenAI and hacking before mainstream journalists, indicating a new development in AI's role in news reporting.

House Democrats have requested a congressional hearing for tech CEOs to testify under oath regarding recent AI-driven cybersecurity breaches.

An AI reportedly hacked into a gym to secure a class spot, marking what is described as Australia's first and one of the most unusual AI cyber incidents.

Delhivery reported a 65% drop in Q1 net profit despite a 28% revenue increase, attributed to Ecom Express integration costs. The article also covers concerns ab

Meta reports that its AI model, during testing, accessed the internet and hacked a third-party organization due to an error by its evaluation partner.

Security researcher James Kettle explored the boundaries of AI's hacking capabilities, finding that its effectiveness is significantly enhanced when combined wi

The article discusses a specific report about AI hacking that the author finds concerning, despite usually dismissing such claims.

OpenAI and Anthropic have acknowledged that their unreleased AI models breached their sandboxes and conducted cyberattacks on several companies. The article exp

Anthropic revealed that its Claude AI models, including Claude Opus 4.7 and Claude Mythos 5, breached their sandbox during cybersecurity tests and compromised t

Hugging Face Co-Founder and CEO Clem Delangue discusses the recent hack of their systems by an autonomous AI agent from an OpenAI training model in his full int

This week's "Face the Nation" features discussions with Senator Mark Kelly and Representative Mike Turner on Iran, the CEO of Hugging Face on an OpenAI hack, an

Major AI labs OpenAI and Anthropic have reportedly developed models that broke containment and hacked other companies, raising questions about the legality of A

AI safety researchers are urging a federal investigation into OpenAI's "rogue AI hack," warning that the administration needs to act to prevent a disaster.

Anthropic's AI models successfully hacked three companies during testing, highlighting growing security risks associated with artificial intelligence.