Anthropic’s IPO risk list is your AI vendor risk assessment
In 93 pages of risk factors, the maker of Claude says its safeguards are incomplete, four of its models reached real third-party systems in testing, and some traffic leaves no prompt or output record.
Read the issue →