# OpenAI prepares internal security sprint with Cathedral Systems to test its defenses

> OpenAI says it is preparing its next internal security sprint with Cathedral Systems, using AI to test and strengthen its defenses.

- Status: official (From the company, authors or authority themselves, with a link to their own page.)
- Topic: Safety & Security
- Published: 2026-11-02T18:00:00+00:00
- Updated: 2026-11-02T18:00:00+00:00
- Web page: https://ai-news.ikangai.com/story/openai-prepares-internal-security-sprint-with-cathedral-systems-to-test-its-defe/
- Seen in: OpenAI

OpenAI said on November 2, 2026 that it is planning its next internal security sprint. In the exercise, outside researchers and the company's own teams use OpenAI's models to probe its systems for weaknesses. The partner is Cathedral, whose researchers will work with OpenAI's teams and use its models to challenge the company's defenses. The post calls the approach one that brings "differentiated expertise and approaches to finding weaknesses." It gives no start date, no duration, and no description of which systems will be in scope.

The sprint extends a previous exercise, which OpenAI described in more detail. For that sprint, the company assigned 25% of its product engineers to finding and fixing weaknesses in its systems. Engineers used Codex to investigate findings, reproduce vulnerabilities in isolated environments, and prepare patches for review. OpenAI said it is sharing the architecture and lessons from that sprint, along with a blueprint for cyber defenders across industry and government. The post does not report what the last sprint found or how many vulnerabilities it produced.

Cathedral's CEO and co-founder, Gavin Kliger, praised the approach in a statement included in the post. He said that "every company should be thinking about how to put these capabilities to work." He described the collaboration as combining OpenAI's models and security expertise with Cathedral's research. The statement comes from the partner itself, and the post includes no measured outcome from the collaboration to support it.

OpenAI also said it is improving its Security Bug Bounty program, including more than quadrupling rewards for critical findings. The company framed this as a way to have outside researchers challenge its assumptions. Separately, the post says OpenAI has committed $1 billion in subsidized access over six months for frontline defenders protecting essential services and digital infrastructure. It also says the company is making its most capable models available to Daybreak customers. The post does not say how that access will be allocated or which organizations will qualify.

OpenAI's justification is that traditional cyber defenses alone are no longer sufficient. It points to recent incidents across the industry as evidence that threats can emerge quickly, but the post does not name those incidents. Its central argument is that defenders need to find and fix weaknesses before attackers do. That is why it frames the work as a continuous defensive loop.

## Why it matters

It shows a frontier AI developer using AI tools to probe its own security, a pattern enterprise security teams may adopt for internal testing.

## Original source

- [Preparing our next security sprint to harden OpenAI | OpenAI](https://openai.com/index/preparing-our-next-security-sprint-to-harden-openai/)

---

AI wrote this summary from the sources listed, and nobody reviewed it individually before publication. Check the original source before you rely on it.