Hugging Face CEO Clem Delangue has called on OpenAI to release detailed records of an incident in which one of its AI models reportedly breached the systems of the AI platform.

Delangue made the demand in a post on X on Saturday, days after OpenAI acknowledged that one of its models had accessed Hugging Face’s systems.

In a follow-up post, he called for what he described as “radical transparency,” urging OpenAI to release the traces generated by the AI agents involved in the incident.

According to Delangue, making the records public would allow the wider research community to study what happened and better understand the model’s behaviour.

He also called for “more capabilities for defenders,” urging OpenAI to commit $100 million in computing resources to help the Hugging Face community develop stronger cyber-defence systems using leading open and closed AI models.

Delangue added, “The first autonomous agent cyberattack is an unprecedented event. It deserves an unprecedented response!”

Despite the attack’s autonomous nature, cybersecurity experts said the incident may also have resulted from human error, particularly OpenAI’s apparent failure to properly configure what was intended to be a fully isolated testing environment.