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:
- Gambian Central Bank says ‘don’t panic’ after data hack
November 18, 2022
The Gambia’s Central Bank says there’s no need to panic after a data hack. Sources allege the hackers managed to access the bank’s most sensitive files, but in a statement the Central Bank said no mission-critical systems were compromised, and that normal operations have continued unabated. It did however say one server was affected, which was promptly ...
- Whoosh confirms data breach after hackers sell 7.2M user records
November 14, 2022
The Russian scooter-sharing service Whoosh has confirmed a data breach after hackers started to sell a database containing the details of 7.2 million customers on a hacking forum. Whoosh is Russia’s leading urban mobility service platform, operating in 40 cities with over 75,000 scooters. On Friday, a threat actor began selling the stolen data on a hacking ...
- Australia: Government considers making cyber ransom payments illegal after Medibank hack
November 13, 2022
It could soon be illegal for companies that fall victim to data breaches to pay ransoms to the hackers. The home affairs minister, Clare O’Neil, confirmed the government was examining whether new laws were needed to stop ransom payments in the wake of the Medibank and Optus data breaches. O’Neil said while short-term successes were needed in ...
- Shangri-La hotel data breach likely had ‘minimal’ impact at Singapore ministerial summit
November 8, 2022
A recent data breach that hit eight Shangri-La hotels is unlikely to have a large impact on foreign government delegates who attended a high-level defence summit in Singapore, which was held at the hotel. Hackers claiming to have instigated the attack apparently have made contact with the hotel chain. Shangi-La Group said Friday it received an ...
- Vodafone Italy discloses data breach after reseller hacked
November 2, 2022
Vodafone Italia is sending customers notices of a data breach, informing that one of its commercial partners, FourB S.p.A., who operates as a reseller of the telecommunications services in the country, has suffered a cyberattack. According to the notice, the cyberattack took place in the first week of September and resulted in the compromise of sensitive ...
- Royal Mail data breach as customers’ information leaked to other users
November 1, 2022
Royal Mail has experienced a data breach where customers have seen the information of others users. A statement on Royal Mail’s Click and Drop status updates website said: “We have been made aware there was an issue affecting Click & Drop that meant some customers could see other customer’s orders. The website is temporarily suspended as “a ...