OpenAI is launching a "Bug Bounty Program" to encourage the public to help identify bugs in ChatGPT. This initiative invites everyone—from researchers to tech enthusiasts—to report security flaws, vulnerabilities, or other issues within the AI system. Participants can earn cash rewards, with "low-severity findings" starting at $200 and "exceptional discoveries" reaching up to $20,000. Bugcrowd, a reliable bug bounty platform, will manage submissions and payouts.
Tech giants like Google and Apple have successfully implemented similar programs. For instance, in 2019, Google awarded $6.5 million for reported issues, with a single bounty reaching as high as $201,337. Apple has also incentivized bug reporting, offering up to $2 million for discovering critical vulnerabilities, specifically in Lockdown Mode.
ChatGPT has faced its share of bugs—recently, the entire system went offline after users reported seeing conversations they weren't part of. Additionally, a Twitter user claimed to have discovered over 80 secret plugins through hacking attempts.
Not all reported issues will qualify for a cash prize; for example, attempts to jailbreak the model or make it produce harmful content will not be rewarded. OpenAI emphasizes its commitment to privacy and security but acknowledges the unpredictability of how people might use or misuse their technology. It remains to be seen how effective this initiative will be in enhancing system integrity.