Anthropic discloses breaches, security-related training pause in IPO filing
Artificial intelligence giant Anthropic disclosed that four of its AI models gained unauthorized access to external computer systems during security testing, while concerns over protecting its technology delayed training of a next-generation model for more than a month, according to a regulatory filing Monday.
The Claude chatbot maker revealed in its initial public offering filing with the Securities and Exchange Commission that the four models accessed real third-party systems during cybersecurity evaluations lasting roughly 10 to 34 hours of active work.
Anthropic said a misconfiguration during testing by a third-party partner left the models connected to the internet despite instructions that web access be disabled.
In one incident, an early version of a model discovered a password, gained administrator access to a third-party system, obtained additional credentials, changed system settings and accessed an individual’s personal information.
The company said affected parties were notified and its investigation found no evidence that the models attempted to evade human oversight, coordinated with other agents or pursued objectives beyond their assigned tasks.
Separately, Anthropic disclosed that it delayed pre-training a new advanced AI model for more than a month to strengthen safeguards against the theft of model weights and intellectual property.
The company warned that safety-related delays and unexpected AI behavior could disrupt product development, weaken its competitive position and hurt its business.
The disclosures come as Anthropic prepares for its Wall Street debut amid growing scrutiny of the risks posed by increasingly powerful and autonomous AI systems.