Come and find loopholes for ChatGPT. OpenAI will receive a maximum reward of 20,000 US dollars.

Beijing, April 12 (Xinhua) artificial intelligence research company OpenAI announced on Tuesday that it will start paying up to $20,000 (140000 yuan) in rewards to people who help the company discover vulnerabilities in artificial intelligence systems, such as the popular chat robot ChatGPT.

ChatGPTOpenAI wrote in a blog post on Tuesday that the company has launched a vulnerability incentive program in partnership with vulnerability reward platform Bugcrowd. Through the program, people can report weaknesses, vulnerabilities, or security problems they find when using their artificial intelligence products. The size of the bonus will depend on the severity of the vulnerability, with a starting bonus of $200, or "low severity discovery", up to $20,000, which requires bug hunters to make "unusual discoveries".

Loophole incentive programs are common in the technology industry, and companies use them to attract people to look for loopholes in their systems. Microsoft, for example, launched a Windows vulnerability incentive program with a maximum reward of $250000. Google has launched a similar program to eliminate vulnerabilities in the Google Play store, with a reward of at least $1000 for each vulnerability.

OpenAI said it launched the program in part because it believes "transparency and collaboration" are the key to discovering its technological vulnerabilities. "this initiative is an important part of our commitment to developing secure and advanced artificial intelligence," Matthew Knight, head of security at OpenAI, wrote in a blog post. "We want your help when we create secure, reliable and trustworthy technologies and services."