On April 12th, Beijing Time, the artificial intelligence research company, OpenAI, announced that it would begin rewarding individuals up to $20,000 (approximately ¥140,000) who assist the company in discovering vulnerabilities in artificial intelligence systems, such as the widely popular chatbot, ChatGPT.
In a blog post on Tuesday, OpenAI stated that it had partnered with the vulnerability rewards platform Bugcrowd to launch a vulnerability rewards program. People can report weaknesses, bugs, or security issues that they have found while using the company’s artificial intelligence products. The size of the bonus will depend on the severity of the vulnerability, starting at $200 for a “low-severity discovery” and reaching up to $20,000 for a “uniquely impactful discovery.”
Vulnerability rewards programs are common in the technology industry, with companies utilizing such programs to attract people to find vulnerabilities in their own systems. For example, Microsoft has previously launched the Windows Vulnerability Rewards Program, with a top reward of $250,000. Google has also launched similar programs in order to address vulnerabilities in the Google Play Store, with each vulnerability being awarded at least $1,000.
OpenAI stated that the reason for launching this program is partly because it believes “transparency and collaboration” are critical to discovering vulnerabilities in its technology. “This initiative is a critical component of our commitment to developing AI that is safe, reliable, and trustworthy,” wrote Matthew Knight, OpenAI’s Security Director, in the blog post. “As we create secure, dependable, and trustworthy technology and services, we want your help.”
