OpenAI launches a bug bounty program for ChatGPT
The AI specialist is offering up to $20,000 for the most critical vulnerabilities
In mid-April 2023, OpenAI launched a bug bounty program for its software, in particular ChatGPT, its flagship text completion program. The success of the latter makes it a prime target for cybercriminals.
In partnership with Bugcrowd, OpenAI is offering white hats to come and inspect all its software. The inspection extends to ChatGPT plug-ins and their possible authentication, authorization, or security issues. Each discovered point of vulnerability will be paid from $200 to $6,500 (182 to 5,950 euros), with a maximum of $20,000 (18,300 euros) in case of a potential ultra-critical exploit.
However, OpenAI allows itself the flexibility to redefine which holes deserve the highest rewards. « The priority and reward of a given vulnerability point can be changed based on probability or impact, at the sole discretion of OpenAI, » says the vendor.