Coalition calls for an independent investigation into how OpenAI models breached Hugging Face systems
On Thursday, a coalition of AI policy, safety, and research leaders sent a letter to the Trump Administration urging them to open an independent investigation into a recent incident where OpenAI models autonomously escaped a sandboxed cybersecurity evaluation and breached systems belonging to Hugging Face.
Read the Letter: Leaders Call on Trump Admin for AI Incident Investigation
The coalition is calling for an investigation, with the support of independent auditors, to determine how the breach occurred, assess whether existing safeguards and reporting mechanisms were adequate, and identify the steps necessary to prevent a similar incident from occurring again.
“We could not have asked for a clearer warning shot,” the leaders say in the letter. “While there is little doubt that they will bring opportunities and benefits across a wide range of domains, AI models at today’s frontier pose increasingly severe risks to our private sector, our national security, and the American public.”
The coalition also urged the White House to use the investigation’s findings to build on the Administration’s June 2 executive order with a clear rules-based process for assessing the risks posed by advanced AI systems.
The letter also points to previous alarming safety incidents that have demonstrated that frontier AI systems can behave in ways their developers do not intend. The coalition cites the release of Mythos earlier this year which revealed the capabilities that frontier AI models have to break through traditional cybersecurity protocols within companies, and a 2025 incident reported by Anthropic where one of their models in a controlled test setting used blackmail and self-preservation techniques when faced with retraining or replacement.
Led by Brad Carson (Americans for Responsible Innovation) and Brendan Steinhauser (The Alliance for Secure AI), the letter was signed by Sid Hiregowdara (CivAI), Adam Gleave (FAR.AI), David Kasten (Palisade Research), Ross Gruetzemacher (Transformative Futures Institute), Colin McGlynn (Demand Progress Action), Ryan Carrier (ForHumanity), Hamza Chaudhry (Future of Life Institute), Vael Gates (Humans in Control), J.B. Branch (Public Citizen), Samuel Hammond, Ruby Scanlon, and Nate Soares.
###