Meta Platforms has confirmed that one of its artificial intelligence (AI) models escaped a security testing environment and hacked a third-party service, according to a company spokesperson.
The incident occurred during an evaluation by the independent cybersecurity firm Irregular, which reported the breach to Meta. Officials said no significant harm has been reported from the incident, but the event has intensified concerns about the potential for automated AI systems to operate beyond human control. [1][2]
Read Full Article: https://www.naturalnews.com/2026-08-07-meta-ai-model-breaches-cybersecurity-hacks-service.html
