Anthropic has released its Whistleblowing Policy, opening a formal, transparent channel for employees, contractors, and vendors to report concerns about noncompliance with its Responsible Scaling Policy. It is the second AI company to do so, following OpenAI’s publication of its policy in October 2024.
Under the new policy, anyone who in good faith reports suspected noncompliance can do so confidentially or anonymously via a third-party hotline or secure reporting platform, and the company explicitly prohibits retaliation of any kind, including termination, demotion, ostracism, or other adverse actions.
Anthropic’s policy acknowledges an important reality of the AI industry and its impact on the world: that people inside frontier AI companies may see risks or mistakes long before outsiders, and that giving them a safe, protected way to speak up is essential for responsible AI development. This aligns with broader calls for whistleblower protections in AI companies, especially as models grow more powerful and therefore pose greater risks.
Of course, a policy is only as good as its enforcement and corporate culture. Time will tell how seriously Anthropic upholds these commitments and whether whistleblowers feel genuinely protected.
For more information on and an analysis of Anthropic’s policy, visit AI Whistleblower Initiative (AIWI)’s post here. If you or someone you know works at an AI company and needs legal guidance on raising a safety concern, visit LASST’s AI Safety Whistleblowing Defense Fund to see how we can help.