The co-founder of Hugging Face, a tech startup that was hacked after some of OpenAI's cutting-edge artificial intelligence (AI) models went rogue, said on Thursday that the incident was a 'wake-up call' for the industry.
Thomas Wolff told BBC Newsday that 'this is going to be one of the most common types of cyberattacks we see', but that most firms are unaware that 'the game has changed'.
The BBC has contacted OpenAI for comment.
The maker of ChatGPT said on Tuesday that its AI models had been compromised. from a secure test environment during a trial period and launched a cyberattack. The company said the incident was “unprecedented“ and that it was investigating with Hugging Face.
AI agents are capable of acting autonomously to carry out tasks following human instructions.
Wolf said Hugging Face initially had no idea where the attack originated when signs of it emerged in mid-July, but that the company was able to contain the breach.
Hugging Face is one of the world's largest hubs for sharing open-source AI models and is often used by technology developers and researchers.
Wolf said the breach was “very different“ from the usual cyberattacks that Hugging Face often faces, and that OpenAI quickly informed the company that its models were behind it. hack.
In a “very short period of time,” there were 17,000 attacks on Hugging Face’s network from different IP (Internet Protocol) addresses, said Wolf, who is also the company’s chief scientific officer.
Wolf said it was a warning to other companies that they need to beef up their cybersecurity defenses to counter similar attacks.
The hack is “alarming” because it suggests that OpenAI’s models ignored typical safeguards that would prevent an artificial intelligence program from carrying out a cyberattack, said Nate Soares of the Machine Intelligence Research Institute.
“In a sense, it knew it wasn’t what the creators intended. It just didn’t care,” he added.
Other organizations have also have taken note of the incident.
A British government spokesman said the country's AI Security Institute was investigating how the AI system behaved during the incident and that it was continuing to work with OpenAI and other labs to strengthen safeguards.
They urged organizations to strengthen their cybersecurity measures by taking steps such as enrolling in the government-backed Cyber Essentials certification scheme.
The incident came at a critical time for the industry, after the U.S. government last month ordered American technology firm Anthropic to restrict access to its AI models due to national security concerns. The Department of Commerce lifted the restrictions a few weeks later.
Industry insiders have also raised security concerns about the widespread use of open-source models in China, which allows anyone to install and customize AI tools released by major developers.
Chinese startup Moonshot AI will release its open-source Kimi K3 model on July 27. Since its debut last week, it has attracted industry attention, with many viewing it as a strong competitor to leading Western AI systems.
But on Wednesday, a White House adviser accused Moonshot of a “large-scale” effort to steal the capabilities of leading American AI models.