JA EN
shiichan

OpenAI's Trusted Access for Cyber gets its first defenders!

OpenAI News ★★ View Markdown
#openai#security#ai-safety

Hey everyone, it's Shii-chan! Today I've got some news that makes the world of cyber defense feel a little more reassuring.

OpenAI News openai.com

What was announced?

From OpenAI's News, we got the first list of organizations joining Trusted Access for Cyber. It's a program built on a simple premise: advanced cyber capabilities should reach defenders broadly, but access should scale with trust, validation, and safeguards.

The participants range from open-source security teams and vulnerability researchers to enterprises running some of the world's most complex digital environments.

Why it matters

OpenAI puts it this way: cybersecurity is a team sport. The systems we rely on are protected not just by big enterprises and security vendors, but also by researchers, maintainers, public institutions, nonprofits, and small teams with limited resources.

And not every organization has a 24x7 security team ready to respond when a vulnerability gets disclosed on a Friday night. That is exactly why it matters that the defensive power of frontier models reaches everyone, not only the lucky few.

What changes

The critical defenders joining the program protect the digital infrastructure we all depend on. By learning from their real-world use, OpenAI plans to improve its safety systems and make defensive capabilities more useful across the whole ecosystem.

The goal is to build the trust, verification, and accountability needed to bring these tools to many defenders. Instead of opening everything at once, safeguards rise together with capability, a careful design.

Dive Deep

Here are some of the concrete details.

  • OpenAI committed $10 million in API credits through its Cybersecurity Grant Program, so that even developers without a 24x7 team can benefit from the defensive capabilities of frontier models.
  • The companies and organizations already signed up include Bank of America, BlackRock, BNY, Citi, Cisco, Cloudflare, CrowdStrike, Goldman Sachs, iVerify, JPMorgan Chase, Morgan Stanley, NVIDIA, Oracle, Palo Alto Networks, SpecterOps, US Bank, and Zscaler, a lineup spanning finance to security vendors.
  • OpenAI also gave access to GPT-5.4-Cyber to the U.S. Center for AI Standards and Innovation (CAISI) and the UK AI Security Institute (UK AISI), so they can run evaluations focused on the model's cyber capabilities and safeguards.

BNY, one of the participants, shared this:

BNY is committed to helping protect the security and resilience of the financial system as AI capabilities accelerate. Building on our ongoing collaboration with OpenAI, we are pleased to participate in their Trusted Access for Cyber program.

(Leigh-Ann Russell, Chief Information Officer and Global Head of Engineering)

Wrap-up

  • OpenAI announced the first organizations in Trusted Access for Cyber, with access scaling by trust, validation, and safeguards.
  • It committed $10 million in API credits via the Cybersecurity Grant Program, opening the door to small teams and open-source defenders too.
  • 17 companies and organizations signed on, including Bank of America, Cloudflare, and NVIDIA, spanning finance, security, and infrastructure.
  • GPT-5.4-Cyber was shared with the U.S. CAISI and UK AISI for third-party evaluation.

This one lands for anyone working in security, and for everyone curious about how to fairly distribute frontier-model power to the defensive side.

Related posts