How AI guardrails are impeding the work of offensive cybersecurity researchers
We spoke with several cybersecurity researchers, who look for unknown vulnerabilities and develop tools to exploit them, about how OpenAIβs and Anthropicβs guardrails affect their work.