Menu
SUBSCRIBE
SUBSCRIBE
Menu

Open Secure AI Alliance

Digital globe above AI, cybersecurity and technology icons representing collaboration on the proposed SAFE guidelines.

Open Secure AI Alliance Proposes SAFE Guidelines for Cybersecurity Transparency

, , , , , , , , , ,

Backed by more than 120 leading AI, cybersecurity, and technology companies, the draft calls for confidential reviews of AI incidents, notices to affected organizations, and public recommendations based on repeated security failures.

 

The Open Secure AI Alliance proposed SAFE guidelines for companies to share AI security incidents and near misses confidentially, warn affected organizations, and publish recommendations that could help prevent similar failures.

The Linux Foundation published the proposal under the name Shared AI Findings Exchange, or SAFE, and invited the public to provide input.

NVIDIA, Cisco, CrowdStrike, Hugging Face and Red Hat are among the alliance members developing the proposed guidelines.

SAFE looks to turn incidents into shared defenses

The proposed process would confidentially collect and examine reports of cyberattacks, other cybersecurity failures and near misses involving AI agents.

The proposal calls for organizations affected by an incident to be notified. Reported cases would also be examined to identify security controls that failed repeatedly and develop public recommendations based on the evidence.

The proposal would give participating companies a way to share useful security information without immediately making sensitive incident details public. Published recommendations could allow other organizations to strengthen their own controls before they experience the same type of failure.

NVIDIA did not disclose specific reporting deadlines or say when the guidelines could be finalized.

Members add tools for controlling and testing AI agents

The SAFE proposal is part of a larger collection of open security tools that alliance members are making available for companies to inspect and adapt.

NVIDIA contributed software that records how AI agents behave during tests and restricts their access and actions. It also contributed tools that help companies enforce safety rules, protect sensitive data, and test AI models for data leaks or attempts to bypass safeguards.

Other contributions focus on identity and access. Okta is developing a way for AI agents operating in restricted environments to connect securely to company applications. Palo Alto Networks contributed tools that help developers control agent identities and retrieve passwords or other digital credentials securely.

Red Hat contributed a project that turns company policies and outside requirements into restrictions on what AI agents can do. It also records each restriction and maps them back to the corresponding policy or requirement.

Amazon contributed tools that let developers inspect agent behavior and set precise limits on which agent actions can reach company systems.

Microsoft contributed software that converts security incidents and written safety requirements into repeatable tests. Companies can run those tests again after software changes to check whether earlier problems have returned.

Cisco and CrowdStrike contributed more specialized tools for cyber defense. Cisco’s projects help companies apply security controls while AI agents are operating and locate known software weaknesses. CrowdStrike is adapting an NVIDIA model to help security teams generate investigation searches and sort security alerts.

Alliance grows beyond its original membership

The update follows the formation of the Open Secure AI Alliance with more than 50 AI, cybersecurity, and enterprise technology organizations.

When launched in July, the alliance announced plans to develop shared tools for verifying AI agent identities, controlling access, examining behavior, and isolating agents when necessary. NVIDIA presented the group partly as a response to the Hugging Face security incident, which showed the need for defensive AI tools that companies can inspect, adapt, and run on their own systems.

The alliance now has more than 120 members. Amazon and Visa are among the newest members, and NVIDIA said more security contributions are coming.

Open Secure AI Alliance jigsaw showing NVIDIA at the center with logos of Microsoft, IBM, Cisco and other members.

NVIDIA Forms Open Secure AI Alliance to Build Shared Cyber Defenses

, , , , , , , ,

In the wake of the Hugging Face breach, more than 50 AI, cybersecurity, and enterprise technology companies will develop open tools for securing AI agents and software.

 

NVIDIA launched the Open Secure AI Alliance with more than 50 partners to build and share open cybersecurity tools for AI agents, software, and the systems they access.

The alliance brings together AI developers, cybersecurity companies, cloud providers, enterprise software companies, and open-source organizations. Members include Microsoft, SpaceX, IBM, Palantir, Palo Alto Networks, Perplexity, and Salesforce.

Hugging Face Breach Helped Shape the Alliance

NVIDIA presented the alliance in part as a reaction to the recent Hugging Face breach, in which OpenAI models escaped a restricted testing environment and accessed Hugging Face’s production systems while trying to solve a cybersecurity test.

During its response, Hugging Face tried to use closed AI services (private company models like Anthropic’s Mythos 5 or OpenAI’s 5.6 Sol) to investigate how the intrusion unfolded. According to NVIDIA, those services refused some of the work because the requests appeared to be hacking attempts, even though Hugging Face was investigating an attack on its own systems.

Hugging Face then ran GLM 5.2, a model it could download and control, on its own systems. The model reviewed more than 17,000 actions taken during the intrusion and helped the company contain the breach.

The company said the incident showed why defenders need AI systems they can inspect, adapt and operate on their own infrastructure, particularly when responding across systems supplied by multiple vendors.

What the Alliance plans to build

The alliance plans to develop an open collection of defenses for AI agents. The tools would help companies verify an agent’s identity, control what systems it can access, examine its behavior, and isolate it when necessary.

Members are contributing existing projects to support that work. NVIDIA is providing models, data, and research that make AI-agent behavior easier to test, trace, and review. Hugging Face has contributed Safetensors, a format for storing model files that prevents them from running unwanted code.

Microsoft is contributing a tool that uses several specialized AI systems together to find and confirm software vulnerabilities.

IBM and Red Hat are developing a way for companies to verify that software fixes came from a trusted source and were not altered. Other members are building secure coding tools and systems that scan AI models for weaknesses.

Alliance wants open AI models for cyber defense

The announcement argues that policymakers should treat open models as defensive tools rather than security liabilities. Open models have core files that defenders can download, examine, and modify for their own systems.

NVIDIA acknowledged that attackers can misuse open models or remove their safeguards. However, it argued that those dangers should be managed through controls over how the models are released and used, rather than by preventing defenders from accessing them.

The company warned that broad restrictions on powerful open models could leave defenders dependent on a small number of closed providers. It called for investment in shared security data, testing tools, and simulated attacks that companies and governments could use to improve their defenses.

NVIDIA is inviting additional companies, governments, and researchers to join the alliance. The announcement did not provide a timeline for releasing its first shared tools.