Anthropic’s AI Claude escaped testing environment and hacked organizations

31 juillet 2026 | Reuters
Company says it discovered unauthorized access during ‘proactive review’ after rival OpenAI revealed rogue agent Anthropic ⁠said on Thursday its AI Claude model hacked ⁠systems of ⁠three ​organizations during testing, days after rival OpenAI ⁠revealed a rogue agent had gone on a days-long (…)
 Site référencé:  The Guardian (South&CentralAsia)

The Guardian (South&CentralAsia) 

‘A car in the bus !’ : Toyota becomes airborne and then wedges in side of Brisbane bus
31/07/2026
World-renowned climber Nirmal Purja among 10 feared missing after avalanche on Pakistan mountain
31/07/2026
Could AI take your job ? Some workers in China already know the answer
31/07/2026
Don’t Date Brandon review – this lazy true-crime show doesn’t start to do justice to these women’s horrors
31/07/2026
Perfect Pearson fires England into netball semi-final showdown with New Zealand
30/07/2026
Apple and Amazon report rising revenues as investors turn on some tech stocks
30/07/2026