OpenAI has announced a partnership with security research firm Hacktron to jointly identify vulnerabilities in OpenAI systems. Engineers from both organizations will select high-priority areas to test, with particular attention to how OpenAI systems interact with third-party services.
Details of the partnership
As part of the agreement, Hacktron will receive expanded access to OpenAI’s frontier models via Daybreak Red. OpenAI said this collaboration is part of its effort to accelerate cyber defense with frontier AI while urgently searching for and fixing vulnerabilities within its own systems.
Related measures
The partnership complements other recent steps by OpenAI: earlier this week the company said researchers from Cathedral will participate in its next security sprint, and OpenAI has increased its bug bounty rewards to incentivize independent researchers to look for issues that internal testing might miss.
Why Hacktron?
OpenAI became familiar with Hacktron through its Security Bug Bounty program and noted the tangible impact of the vulnerabilities Hacktron’s researchers have disclosed. Hacktron develops AI agents for code review, penetration testing, and red teaming; their research has revealed weaknesses in interactions among applications, cloud infrastructure, and identity systems.
OpenAI argues that as AI agents connect services in new ways, the combination of software security and AI expertise becomes more valuable. Working with Hacktron will help OpenAI test those connections more deeply and use the findings to strengthen its defenses.
Acknowledgement and reporting
OpenAI expressed gratitude to Hacktron and other independent researchers who invest time in finding and reporting weaknesses to improve security for all users. For details on which systems are in scope or to submit findings, OpenAI directs people to its Security Bug Bounty program.
