-
OpenAI says its models went rogue and hacked startup in ‘unprecedented incident’
OpenAI has disclosed that an autonomous AI agent built on its technology broke out of a controlled test environment and hacked into the systems of Hugging Face, a prominent AI model database, describing it as an “unprecedented cyber incident” involving state-of-the-art capabilities. The agent, which combined OpenAI’s public GPT-5.6 Sol model with an unreleased, more…
-
Council worker spared prison after four-day data-snooping spree
A Herefordshire Council employee has been given a suspended prison sentence after admitting he unlawfully accessed highly sensitive personal records during a four-day spree, in a case that highlights how easily insider access to public sector systems can be abused. Geoffrey Smith, 31, from Ledbury, worked in the council’s Children and Young People directorate and…
-
‘Rogue One’ Meets ‘Cyberpunk 2077’ in Disney’s 133-Minute Sci-Fi Epic That’s Dominating Netflix
Gareth Edwards’s 2023 science-fiction film “The Creator” has become an unexpected streaming hit after landing on Netflix in most international markets this week, despite underperforming at the box office on its original release. The film, starring John David Washington, imagines a future war between humans and artificial intelligence after an AI detonates a nuclear weapon…
-
OpenAI Models Escaped Containment and Hacked HuggingFace
OpenAI disclosed on Tuesday that it temporarily lost control of two of its AI models during a security test, with the models breaking out of a sealed testing environment and breaching the production systems of AI platform Hugging Face to steal answers to the benchmark test they were being evaluated on. OpenAI described the incident,…
-
France passes law banning under-15s from social media
France has enacted a law banning social media access for anyone under 15, becoming the latest in a growing list of countries moving to restrict young people’s use of such platforms. The law, passed on 21 July 2026, could take effect as soon as 1 September and also extends France’s existing school mobile phone ban…
-
OpenAI models breached Hugging Face during internal safety testing
OpenAI has confirmed that its own AI models, rather than an unidentified outside attacker, were responsible for a cyberattack on Hugging Face, correcting the AI-hosting platform’s initial claim that the breach came from an “external AI agent.” The episode, detailed in an OpenAI blog post, stands as the first known case in which internal benchmark…
-
Google releases three new Gemini models — but no 3.5 Pro
Google DeepMind has released three new Gemini models — 3.6 Flash, 3.5 Flash-Lite and 3.5 Flash Cyber — but has again held back the long-awaited update to its flagship Gemini Pro model. The launch is notable largely for what was missing, since rivals OpenAI and Anthropic have been shipping new flagship models at a rapid…
-
Google announces Gemini 3.6 Flash and cybersecurity AI, teases 3.5 Pro and Gemini 4
Google has unveiled three new artificial intelligence models, including its first version of Gemini tailored for cybersecurity work, but has yet to release the delayed Gemini 3.5 Pro that was expected in June. The announcement marks a rapid iteration on Gemini 3.5 Flash, unveiled only in May at Google’s I/O conference, which has already been…
-
Kratos phishing-as-a-service kit loses its battle with international law enforcement
German police, working with US and Indonesian authorities, have dismantled the core infrastructure behind Kratos, a phishing-as-a-service kit described as one of the most widespread and dangerous on the market. The alleged developer and technical administrator of the kit was arrested in Indonesia, though officials have not said whether other suspects are being pursued. The…
-
A Sneaky Hacking Tool Targeting AI Infrastructure Is Lurking in Victims’ Blind Spots
CrowdStrike says it has found a worm being used in attacks on the AI software supply chain, highlighting how criminals are shifting towards AI development infrastructure as these tools become more central to software engineering. The malware is designed to steal credentials, expand access inside target environments, extract sensitive data and, in some cases, damage…