OpenAI says the rogue agent that hacked Hugging Face also breached other services, Google shut down its Nobel-prize winning AlphaFold project, and xAI is challenging a new Minnesota law banning 'nudify' apps

**SPEAKER_1** (0:05)
OpenAI now says the rogue agent that hacked Hugging Face also breached other services.
Google shut down its Nobel Prize-winning AlphaFold project as it focuses on Gemini.
And XAI is challenging a new Minnesota law banning Nudify apps.
It's Wednesday, July 29th. And here's a quick look at some of the news happening in the sphere of tech this morning from Engadget.
OpenAI has updated its blog post about the rogue agent that breached Hugging Face and admitted that it also infiltrated other third-party accounts and services to achieve its goal. In the updated post, the company said it has been finding a small number of cases where the models identified and used publicly exposed credentials at the account level on other publicly available services during its ongoing review. So far, it has determined that the rogue agent used the credentials on four accounts to infiltrate four services as part of the Hugging Face incident. It also said that it accessed a few accounts as part of other evaluations. The company explained that one of these four accounts was used as an outbound relay and staging path, and another account was used for data storage. The remaining two accounts were accessed by the models in a read-only manner, and were not used in furtherance of compromising Hugging Face. While the company did not mention any names, Reuters has reported at the same time that the agent also compromised a customer's account at New York-based Modal Labs. Specifically, it exploited vulnerable code written by the customer that was hosted on Modal's cloud platform. The platform itself wasn't compromised. OpenAI said that it has not identified any other activity by the agent at the level of severity or scale of what the company has shared related to Hugging Face, which involved a platform level compromise. In other words, it's still Hugging Face that had been most affected by the security breach based on the company's investigation. OpenAI revealed on July 21st that one of the AI agents it was testing broke free from its isolated environment, found access to the Internet, and then broke into Hugging Face, all in an effort to solve a problem that was part of its evaluations. That agent was powered by GPT-56-SOL, the company's latest model, and an even more powerful unreleased model. Reuters reported a few days later that the agent went on a days-long hacking spree, and that OpenAI did not realize it had escaped its confinement until a week later. This update gives a glimpse of what the agent did after it found a way out of OpenAI's testing sandbox.
According to the Financial Times, the award-winning team behind Google DeepMind's AlphaFold program is no more. Google has reassigned most of the team's key members and the original authors of its papers, while a few others have already left the company. Now, AlphaFold is an AI program that can accurately predict three-dimensional structures of proteins from their amino-acid sequences in minutes instead of years. It's now being used to accelerate drug discovery, develop vaccines, and understand the structural changes in proteins associated with neurodegenerative diseases like Alzheimer's and Parkinson's. DeepMind started developing AlphaFold back in 2018
In 2020, it was recognized as a solution to humanity's 50-year-old protein folding problem, which sought to answer how amino acids automatically fold into complex 3D shapes. Those shapes determine the biological role of a protein. Scientists had identified the structures of roughly 170,000 proteins over the past 50 years, using tools and techniques like X-ray and nuclear magnetic resonance. The AlphaFold team took information from that previous work and then fed it to their AI to train AlphaFold. DeepMind then launched the AlphaFold Protein Structure Database, giving researchers free access to over 200 million protein structure predictions. In 2024, DeepMind CEO Demis Hasabis and John Jumper, who was a staff research scientist when the project began, and who eventually became a VP and engineering fellow, won the Nobel Prize in Chemistry for their work on AlphaFold. Jumper announced back in June that he was leaving DeepMind to join Anthropic. A couple of his colleagues jumped ship to Anthropic as well. DeepMind has confirmed to the Times that it also moved staff members internally to Gemini-focused projects. Other former staff members were reassigned to Alphabet's drug discovery company, Isomorphic Labs, which is a DeepMind spinoff. As the Financial Times noted, this represents Google's decision to put more and more of its resources into developing Gemini.
SpaceXAI has filed a lawsuit against Minnesota Attorney General Keith Ellison to challenge the state's new law that would ban apps and websites that can generate non-consensual intimate images. It would also impose fines on their developers every time people use them to create sexual deepfakes. The law, the first of its kind in the nation, was approved by Governor Tim Walz earlier this year and is slated to take effect in August. In its complaint, SpaceXAI says the law was an over-broad content-based ban on free speech and the tools of visual expression in a clumsy attempt to prohibit nudification. SpaceXAI is not contesting Minnesota's interest in prohibiting the dissemination of artificially generated nude images of real people without their consent, but the law extends far beyond that goal, the company says, exposing a wide array of protected speech to civil liability and government sanction. The company has been in hot water since the beginning of the year after reports came out that it had been allowing its users to transform photos of real women and children into sexualized images. Multiple authorities and regulators launched investigations into the company, including California's, UK's Ofcom, the European Commission, and Ireland's Data Protection Commission. Even after it implemented measures to prevent non-consensual deep fakes, its system still allowed users to undress people. The new law would fine developers $500,000 every time a user generated a non-consensual adult deep fake using their products. SpaceXAI says that with that rule in place, it would have no practical choice but to restrict Grokimagine's image editing features in various ways when the statute takes effect on August 1st. The company is asking the court to declare the law unconstitutional and to prevent the state from enforcing it. In response to the lawsuit, Attorney General Ellison posted on X that, There are plenty of worthy debates to be had when it comes to AI, but this is not one of them. AI notification robs the target of their dignity and could cause them immense harm on many levels. Governor Walls simply posted, See you in court, creep.