Anthropic says Claude hacked three online systems in cyber-testing incidents
( July 30, 2026, 22:00 GMT | Official Statement) -- MLex Summary: Anthropic has said its advanced AI models have now escaped their test environment three times and gained unauthorized access to third-party organizations. Following an incident at rival OpenAI a week prior, Anthropic said it checked its own Claude system and found it had also “gone rogue” during cybersecurity simulations. Due to a “misunderstanding” in a prompt, Claude gained access to the internet and targeted real systems. Anthropic said that going forward, it would improve evaluation environments and its monitoring of results.Statement follows....
Prepare for tomorrow’s regulatory change, today
MLex identifies risk to business wherever it emerges, with specialist reporters across the globe providing exclusive news and deep-dive analysis on the proposals, probes, enforcement actions and rulings that matter to your organization and clients, now and in the longer term.
Know what others in the room don’t, with features including:
- Daily newsletters for Antitrust, M&A, Trade, Data Privacy & Security, Technology, AI and more
- Custom alerts on specific filters including geographies, industries, topics and companies to suit your practice needs
- Predictive analysis from expert journalists across North America, the UK and Europe, Latin America and Asia-Pacific
- Curated case files bringing together news, analysis and source documents in a single timeline
Experience MLex today with a 14-day free trial.