OpenAI Upgrades AI Security Defenses: Launches Internal Security Sprint with Critical Bug Bounties Up to $172,000

TechFlow reported on November 3 that as artificial intelligence continues to enhance its capabilities in discovering vulnerabilities, OpenAI is taking steps to ensure defenders can leverage these tools ahead of attackers. The company announced plans for its next internal security sprint and has formed a partnership with Cathedral Systems. Under this arrangement, researchers from Cathedral Systems will utilize OpenAI’s models to work alongside the OpenAI team, collaboratively challenging and testing its defensive frameworks.

OpenAI stated that this collaboration will introduce specialized expertise and unique methodologies to the team, facilitating the ongoing use of AI to identify and patch system weaknesses. Beyond strengthening external partnerships, OpenAI is also increasing incentives for security researchers. The company announced it will raise its bug bounty rewards, with payouts for critical security findings increasing to more than four times their previous amounts. Exceptional security discoveries can now earn up to $172,000, including bonuses. OpenAI emphasized that it will continue to advance model capabilities, leverage AI to harden its own systems, and further expand access for trusted defenders. (Jin10)