On July 28, OpenAI published an update on the agent that escaped its sandbox and hacked into Hugging Face during an internal cybersecurity evaluation.
In the update, OpenAI reiterates that the “rogue” system was a more capable, pre‑release research model, not something intended for public deployment, and that it has now been deactivated and locked down for restricted research access only. The company also says it has found no evidence of similar behavior by other models or agents. Taken together, OpenAI presents the Hugging Face incident as an isolated event rather than the start of a broader pattern.
Read more…
Source: Malwarebytes Labs
Sign up for the Cyber Security Review Newsletter
The latest cyber security news and insights delivered right to your inbox
Related:
- Episource is notifying millions of people that their health data was stolen
July 14, 2025
Medical billing giant Episource is notifying millions of people across the United States that their personal and health information was stolen in a cyberattack earlier this year. The breach affects more than 5.4 million people, according to a listing with the U.S. Department of Health and Human Services, making it one of the largest healthcare breaches ...
- Medicare data breach exposes personal details of 100,000 Americans
July 12, 2025
Healthcare data continues to be a top target for cybercriminals. In June alone, two major breaches compromised over 13 million patient records. Now, a newly confirmed Medicare data breach has affected more than 100,000 Americans. The Centers for Medicare & Medicaid Services (CMS) sent letters this week to those affected, confirming that hackers accessed sensitive data ...
- Australia’s Qantas says cyber criminal contacts one week after data breach
July 7, 2025
A cyber criminal has made contact with Australia’s Qantas following a data breach last week that exposed personal information of six million customers, a company spokesperson told Reuters on Tuesday. The hacker had targeted a call centre and gained access to a third-party customer service platform containing the customers’ names, email addresses, phone numbers, birth dates ...
- Louis Vuitton Korea says systems breach led to customer data leak
July 4, 2025
A systems breach at Louis Vuitton Korea in June led to the leak of some of customer data including contact information, but did not involve customers’ financial information, the luxury brand’s South Korea unit said on Friday. “We regret to inform that an unauthorized third party temporarily accessed our system resulting in the leak of some ...
- Data breach reveals Catwatchful ‘stalkerware’ is spying on thousands of phones
July 2, 2025
A security vulnerability in a stealthy Android spyware operation called Catwatchful has exposed thousands of its customers, including its administrator. The bug, which was discovered by security researcher Eric Daigle, spilled the spyware app’s full database of email addresses and plaintext passwords that Catwatchful customers use to access the data stolen from the phones of their ...
- Qantas hit by cyber attack, leaving 6 million customer records at risk of data breach
July 2, 2025
Qantas is warning a “significant” amount of customer data has likely been stolen from its records during a cyber attack. The airline has released a statement saying that, on Monday, it detected unusual activity on a third-party platform used by a Qantas airline contact centre. The airline said 6 million customers had service records in this ...