Anthropic expands AI access for security
Anthropic is expanding a program allowing cybersecurity professionals to test its most powerful AI models with fewer safeguards. The initiative follows the discovery of over 100000 software vulnerabilities.
Anthropic is expanding access to its most powerful AI models for vetted security professionals, offering reduced safeguards after early testing helped uncover 129,000 verified vulnerabilities. The move builds on the Project Glasswing initiative, which launched earlier this year to secure critical software infrastructure. Over 33,000 of the discovered flaws have been rated critical or high severity, according to company data released on Tuesday.
