After the egg hunt, Internet users can perform a bug hunt. On Tuesday, the OpenAI company opened its bug detection program. The goal is to secure ChatGPT, its AI-powered chatbot. To motivate participants to search in all corners of the tool, tempting rewards are offered.
Based on the BugCrowd platform specialized in reporting computer errors, OpenAI indicates that the investigation refers to ChatGPT itself, but also to its API (programming interface) and its integration into certain third-party applications.
Two priority failures
Reported errors fall into four categories. Associated bounties range from $200 to $500 for minor bugs. Subsequent levels go up to $1,000, $2,000, and $6,500 for first-order errors. Flaws identified as critical by OpenAI alone will likely net hackers $20,000.
So far, 23 reports have been offered through BugCrowd. Among them, two are issues identified as top priorities. Netizens who pointed out these flaws received $2,000 and $6,500.
In August 2022, SpaceX had used a similar device to sanitize the code for its Starlink service. This bug hunt was started after a hack of the satellite internet access system. All it took was a homemade device that cost just $25 for a hacker to use the service for free.
Source: BFM TV
