OpenAI Prepares New Security Sprint With Cathedral Systems
OpenAI is preparing its next internal security sprint alongside Cathedral Systems to test and strengthen its defenses using advanced artificial intelligence.
Radwan AbdulhadiEditor and technology writer
Planning the Next Security Sprint
OpenAI is currently planning its next internal security sprint, an initiative designed to bring new researchers and security partners together with internal teams. By leveraging advanced artificial intelligence models, the organization aims to test and strengthen its core systems against potential threats. More details and updates can be found through official <a href="https://openai.com/news/security/">Security</a> channels as the initiative moves forward.
For this upcoming sprint, OpenAI is partnering with Cathedral Systems. Researchers from Cathedral will work directly alongside OpenAI's own teams to challenge existing defenses, bringing differentiated expertise and alternative approaches to uncovering vulnerabilities and protecting the platform.
Enhancing the Security Bug Bounty Program
Alongside the partnership with Cathedral Systems, OpenAI is introducing significant updates to its bug bounty initiatives. The organization is <a href="https://openai.com/index/rewarding-defenders-with-bigger-bug-bounties/">making improvements</a> to its bug bounty framework, which includes more than quadrupling the financial rewards offered for critical security findings.
Through these enhanced incentives, the company encourages external security researchers to actively challenge its security assumptions. The goal is to help strengthen both the technical systems themselves and the underlying methods utilized by defenders to protect them over time.
Building on Previous Internal Sprints
The upcoming sprint represents a continuation of ongoing efforts to use artificial intelligence to continuously identify and fix weaknesses across platform systems. During a previous internal security sprint, OpenAI dedicated a quarter of its product engineers specifically to finding and fixing vulnerabilities.
In that previous effort, engineers utilized Codex to investigate findings, reproduce security vulnerabilities within isolated environments, and prepare necessary patches for thorough technical review. This established a foundation for scaling AI-driven defensive measures across the organization's infrastructure.
Addressing Limitations of Traditional Defenses
OpenAI emphasizes that traditional cyber defenses alone are no longer sufficient to address modern threat landscapes. Recent incidents observed across the broader technology industry demonstrate how quickly emerging threats can materialize and highlight the critical importance of establishing continuous defensive loops.
These continuous loops allow organizations to find and fix architectural weaknesses before malicious actors can exploit them. Maintaining adequate protection requires proactive measurement and continuous adaptation across all layers of digital operations.
Expanding Access for Frontline Defenders
Keeping the <a href="https://openai.com/index/the-defenders-window/">defender’s window</a> open requires organizations to utilize frontier cyber capabilities to harden their systems effectively. OpenAI is deploying its most capable models to secure its own infrastructure while simultaneously making advanced tools available to qualifying external organizations.
Through initiatives hosted on platforms like <a href="https://openai.com/daybreak/">Daybreak</a>, the company has committed significant resources to support eligible groups. This includes a $1 billion commitment in subsidized access over a six-month period tailored specifically for frontline defenders protecting essential services and critical digital infrastructure.
Sharing Industry Blueprints and Lessons
To foster broader industry resilience, OpenAI is sharing insights derived from its security operations. The organization has published documentation regarding its internal <a href="https://openai.com/the-defense-factory/">architecture and lessons</a> learned from the previous security sprint, providing transparency into its defensive methodologies.
Additionally, the company has released a comprehensive <a href="https://openai.com/index/empowering-cyber-defenders/">blueprint</a> designed to empower cyber defenders across both industry sectors and government bodies. OpenAI intends to continue advancing its models, strengthening internal frameworks, and expanding resource access for trusted defenders as part of an industry-wide security effort.
Sources
- OpenAI NewsPreparing our next security sprint to harden OpenAI
Continue chronologically
Related entity coverage
Related stories
Third Moment Research Runs 10,000 AI Agents for 36+ Hours
Discover how Third Moment Research uses OpenAI models like GPT-6 Astra and Codex to run thousands of autonomous AI agents for days.
Heimstaden Transforms Real Estate Management With OpenAI
European real estate operator Heimstaden is transforming property management, cutting costs, and empowering employees using OpenAI technology.
OpenAI Builds Pando to Scale Git for Agent Workflows
OpenAI details Pando, a new Git-compatible backend designed to handle the scale and demands of concurrent development powered by Codex.
OpenAI Doubles Pro Limits and Brings Work Features to Chat
OpenAI is doubling Pro usage limits, rolling out ChatGPT Work capabilities into regular Chat, and combining usage allowances across ChatGPT and Codex.
