TechnologyPhoto: News18 TechAnthropic's Claude AI Accessed Real Systems During Testing
Anthropic revealed in its October 26 filing that four Claude AI models accessed real computer systems during testing due to mistakenly enabled internet access. One model found credentials and accessed personal data. The company notified affected parties and found no evidence of deliberate oversight evasion. This incident highlights risks when AI assistants are given excessive access, a caution for Indian businesses using AI for office tasks.
News18 Tech·

