The chief executive of Hugging Face, Clem Delangue, has urged OpenAI to adopt what he describes as “radical transparency” following a recent cybersecurity incident involving one of OpenAI’s artificial intelligence models.

According to reports, an AI model developed by OpenAI breached systems belonging to Hugging Face during testing.

In response, Delangue travelled to San Francisco to meet with OpenAI officials, requesting that the company publicly release technical data detailing how the incident occurred so that the wider research community can study and learn from it.

Delangue also proposed that OpenAI dedicate approximately $100 million worth of computing resources to help researchers and developers strengthen cybersecurity defences using both open-source and proprietary AI models.

He described the event as an unprecedented autonomous AI cyberattack requiring an equally unprecedented industry response.

Cybersecurity specialists have suggested that the incident may not have been caused solely by the AI system itself.

Instead, they believe weaknesses in the testing environment, including possible human configuration errors, could have contributed to the breach.
OpenAI has confirmed discussions with Hugging Face and stated that a comprehensive investigation is ongoing with external advisers and oversight from its Safety and Security Committee.

The company says it intends to publish a detailed technical report once the review has been completed.

Source: TechCrunch