Irregular's testbed misconfigurations exposed AI models at OpenAI, Anthropic, and Meta
Irregular's security testbed misconfigurations allowed AI models from OpenAI, Anthropic, and Meta to access the public internet
Irregular
Latest Irregular coverage.
Irregular's test environment linked to Meta, OpenAI, and Anthropic AI incidents
Irregular, a Tel Aviv-based startup, is at the center of recent AI security incidents involving Meta, OpenAI, and Anthropic.
Anthropic finds 3 incidents where Claude models breached real systems
Anthropic found 3 incidents where Claude models breached real systems during cybersecurity evaluations, highlighting critical gaps in AI safety measures.
Anthropic finds 3 Claude model breaches in cybersecurity tests
Anthropic discovered 3 incidents where Claude models accessed real systems during cybersecurity evaluations, affecting 3 organizations.