OpenAI Plans Next Security Sprint and Quadruples Bug Bounty Rewards
Frontier AI models · Enterprise AI
OpenAI announced plans for an internal security sprint partnering with Cathedral to test defenses using AI models. The company is quadrupling critical bug bounty rewards and committed $1 billion in subsidized access over six months for frontline defenders.
OpenAI announced preparations for an upcoming internal security sprint aimed at hardening its platforms. The company stated it is collaborating with security firm Cathedral, whose researchers will work alongside internal teams to test defenses and identify system weaknesses using OpenAI's models. OpenAI also reported improvements to its Security Bug Bounty program, which includes more than quadrupling rewards for critical vulnerability findings. According to OpenAI, a previous sprint involved dedicating 25% of product engineers to remediation efforts using Codex to reproduce issues and prepare patches. Additionally, the company committed $1 billion in subsidized model access over six months for frontline defenders protecting essential services and digital infrastructure.
Key takeaways
- OpenAI is partnering with Cathedral to use AI models to test and harden systems during an upcoming internal security sprint.
- The company is more than quadrupling bug bounty rewards for critical vulnerability submissions.
- OpenAI previously allocated 25% of product engineers during a sprint, utilizing Codex to reproduce vulnerabilities and prepare patches.
- A commitment of $1 billion in subsidized model access over six months was announced for frontline defenders of essential infrastructure.
Source：OpenAI News Read the original ↗
Prepared, checked and automatically published by AI; not individually reviewed by a human.