OpenAI Is Rewarding Up To $20K To People Who Discover Glitches In ChatGPT

Spotted a technical inconsistency in your conversation with ChatGPT? OpenAI is giving users a chance to pick its artificial intelligence systems’ brains—and be rewarded for it—with its new bug bounty program.

The Microsoft-backed AI platform, renowned for developing revolutionary tools like ChatGPT and DALL-E, is doling out as much as US$20,000 to people who flag up weaknesses, bugs, or security issues in its products. OpenAI is partnering with bug bounty platform Bugcrowd, which will oversee the submission and reward process, to power this initiative.

OpenAI is accepting discoveries of “low-severity” glitches, with cash rewards starting from US$200, to “exceptional discoveries,” which could yield the maximum reward of US$20,000.

Eligible discoveries include errors in OpenAI’s APIs and cloud infrastructure, as well as evidence of data exposure.

However, content issues, in which you get the bot to “hallucinate” details, are out of scope in the reward program and should be reported directly to OpenAI.

Participants must also promise to keep vulnerabilities confidential until they are authorized by OpenAI to make them public.

In its blog post, OpenAI explains that “transparency and collaboration” are integral to tackling the complexities of AI systems.

“This initiative is an essential part of our commitment to developing safe and advanced AI,” explains head of security Matthew Knight. “As we create technology and services that are secure, reliable and trustworthy, we would like your help.”

[via Bloomberg and Forbes, cover illustration 274809022 © Irina Batiuk | Dreamstime.com]