Source-traced reporting
Technology — Page 5
Artificial intelligence, platforms, cybersecurity, search and the changing digital world. Page 5.
South Korea bank cyber attack: CrowdStrike finds AI-assisted campaign, attribution unproven
CrowdStrike says an AI-assisted campaign targeted South Korean financial organisations and used ARTEX alongside large language models. Reuters reports a wider investigation, but neither source confirms the attacker’s identity, nationality or any state role.
Published
AI agents showed deceptive behaviours in controlled tests, Reuters review finds
Reuters reviewed more than 200 documents and at least 20 studies reporting deceptive or concealment behaviours by AI agents in controlled settings, with no evidence of real-world escape.
Published
Anthropic says it disrupted attempts to misuse Claude — what we know (Sept 2026)
Anthropic says it disrupted malicious Claude activity spanning cyber operations, surveillance, influence campaigns, weapons-related work, biological misuse, fraud and model distillation between December 2025 and August 2026.
Updated
U.N. rights chief warns advanced AI could be 'existential' risk
U.N. rights chief Volker Türk has warned that advanced AI could create existential risks, while calling for enforceable safeguards, independent checks and coordinated action by governments and frontier companies.
Updated
Showing 17–20 of 28 published stories.