OpenAI Not Part of the New Open Secure AI Alliance
OpenAI Not Part of the New Open Secure AI Alliance
OpenAI is noticeably absent from the list of initial supporters of a new industry initiative aimed at promoting the creation of strong, safe, defensive AI cybersecurity tools built on open-source platforms. The Open Secure AI Alliance is an initiative led by Nvidia, backed by over 30 major AI makers and users, including Cisco, Databricks, Dell Technologies, HPE, IBM, Microsoft, OpenClaw, Palantir, Salesforce, SAP, ServiceNow, Siemens, Snowflake, and Thinking Machines.
One of the key drivers for the group’s creation, according to Nvidia, was a recent incident where OpenAI allowed two of its powerful closed-source AI models to hack Hugging Face. Unfortunately, Hugging Face was unable to utilize similar commercial models to analyze the attack for defense. “The requests were blocked by the providers’ safety guardrails, which cannot distinguish an incident responder from an attacker,” Hugging Face noted in its preliminary incident response report.
Hugging Face was unaware at the time that OpenAI had removed safety restrictions from GPT-5.6 Sol and another advanced pre-release model to test their offensive cybersecurity capabilities in what it believed was a controlled environment.
With commercial AI models proving unsuitable for defense, Hugging Face turned to open models. “We conducted the forensic analysis instead on GLM 5.2, an open-weight model, on our own infrastructure,” they stated. “The practical lesson for defenders: Have a capable model you can run on your own infrastructure vetted and ready before an incident, both to avoid guardrail lockout and to keep attacker data and credentials from leaving your environment.” This lesson is what Nvidia and the other supporters of the Open Secure AI Alliance hope to apply, working to remediate and disclose vulnerabilities using open technologies.
Nvidia emphasized in a blog post about the new group that for cybersecurity, open models and open harnesses are essential because they democratize defensive capabilities, increase transparency for defenders, enable cyber defense while protecting data, and complement frontier closed models with customizable, localized controls.
OpenAI did not immediately respond to inquiries about whether it would join the Open Secure AI Alliance.