OpenAI Doubles Rewards to Secure GPT-5.6 Bio

Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisis ton rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
A Reward Program for the Security of Biological AIs
As part of an initiative aimed at enhancing the security of artificial intelligences applied to biology, OpenAI has decided to evolve its bug bounty program related to GPT-5.5 Bio. This program, now known as the OpenAI Bio Bounty Program, is becoming a permanent private program. The main objective of this initiative is to focus on universal jailbreaks, which represent vulnerabilities that allow bypassing the biosafety measures predefined by OpenAI for its most advanced models, starting with GPT-5.6 and subsequent versions.
Significant Increase in Rewards
OpenAI has also announced a substantial increase in the rewards offered under this program. The bounty for discovering a universal jailbreak within the OpenAI Bio Bounty Program has now been doubled, rising from $25,000 to $50,000. This increase applies to the GPT-5.6 and GPT-5.5 models. Additionally, more modest rewards may be granted for partial discoveries, at OpenAI's discretion.
Timeline and Scope of the Program
The reward program for GPT-5.5 Bio will continue to adhere to the initially defined scope, with testing continuing until July 27, 2026. After this date, the program will focus exclusively on GPT-5.6. OpenAI is committed to regularly informing researchers of any future changes to the program's scope.
Simplified Application Process
Individuals interested in this program must submit a brief application, including their name, affiliation, and experience. Once accepted, candidates will be integrated into the dedicated platform for biological bug bounties. It is important to note that former participants in the GPT-5.5 Bio Bounty Program will not need to submit a new application.
How to Participate
To participate, candidates must fill out a short application including their name, affiliation, and experience. Accepted candidates must also have a ChatGPT account and sign a non-disclosure agreement (NDA).
OpenAI invites anyone interested in contributing to the security of cutting-edge artificial intelligence to apply now. Furthermore, those interested in other aspects of AI security can inquire about the Safety Bug Bounty and Security Bug Bounty programs offered by OpenAI.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.