Responsible Disclosure

A critical practice for AI Red Teamers is responsible disclosure: privately reporting discovered AI vulnerabilities (e.g., a successful jailbreak, data leak method, or severe bias) to the model developers or system owners, allowing them time to remediate before any public discussion, thus preventing malicious exploitation.

Learn more from the following resources:

• @[email protected] Policy
• @guide@Huntr Guidelines
• @policy@Google Vulnerability Reward Program (VRP)