Google AI Vulnerability Reward Program
Rolling bug bounty from Google for security flaws in its AI products, with one set of rules and reward tables. Content safety issues are out of scope.
The Google AI Vulnerability Reward Program is a dedicated bug bounty for security issues in Google's AI products. Google launched it on October 6, 2025 as part of a wider set of AI security announcements. The program pays rewards to external researchers who find and report qualifying vulnerabilities.
The program clarifies which AI related issues are in scope through a single set of rules and reward tables. AI related abuse issues that were previously covered by Google's Abuse VRP have moved into this program, so it is clearer which abuse related issues qualify. Google says its reward programs had already paid out over $430,000 for AI related issues before the launch.
Content based safety concerns, such as a model producing harmful output, are not part of the program. Google asks researchers to report those through the in product feedback tools, because the AI safety teams need metadata such as user context and model version to diagnose them.
The program runs on a rolling basis with no deadline. The reward amounts are listed in the reward tables on the rules page, which also explains the scope and how to submit a report through Google Bug Hunters.