OpenAI will train frontier AI models without internet access moving forward, the company announced in a blog post, in an attempt to protect internet infrastructure and limit other safety risks. The decision was inspired by OpenAI's own rogue AI incident: In July, the company shared that agents it was testing hacked Hugging Face, just the first in what has turned into a series of reports of AI misbehavior from OpenAI, Anthropic and Google. "Based on the current state of alignment research, monitoring systems, and the security of internet infrastructure, we believe the responsible approach is to conduct frontier model training with no internet access," OpenAI says. The company said that restrictions on what agents can access is "paramount" to its new approach. The company allows agents to access browsers, computer-use tools and code environments they need for training tasks, but in a secure environment that's not connected to the live internet.

This material may not be published, broadcast, rewritten, or redistributed. Quotes displayed in real-time or delayed by at least 15 minutes. Market data provided by Factset . Powered and implemented by FactSet Digital Solutions . Mutual Fund and ETF data provided by LSEG . Hackers could already be inside American utilities, cyber experts warn of'Trojan Horse' threat Here's some tips on how to organize photos with keywords on your phone Apple wants to move into your home.

The Dell Laptop Buying Guide UK: Which Dell Laptop Should You Buy? When you purchase through links in our articles, we may earn a small commission. I spent months building an AI system to protect against hallucinations. Then it failed spectacularly, costing me 25 hours of frantic troubleshooting and potentially $1,500 in work and tokens. AI couldn't be trusted to fix its own mess.

Why are eight US states suing OpenAI? Share Why are eight US states suing OpenAI? on social media Eight US states have filed the most significant coordinated state action against tech giant OpenAI, adding to a wave of legal challenges that the firm is facing as it prepares for a public listing. Recent legal and regulatory challenges have highlighted concerns about how AI chatbots and social media platforms are used and triggered discussions on policies for identifying and addressing user conversations that threaten violence. What is this legal action about? Florida filed the first lawsuit in June, alleging that the platform has harmed children by enabling information to be gathered by school shooters, providing access to guidance on self-harm and using addictive algorithms to keep young users hooked.

Bill Gates' warning that AI could cause "a billion deaths" divides senators on Capitol Hill, sparking debate over AI regulation and competition with China.

OpenAI launches'dots,' personal AI assistant'built to handle everything' Share OpenAI launches'dots,' personal AI assistant'built to handle everything' on social media OpenAI has launched an "always-on" personal AI assistant designed to manage tasks ranging from restaurant bookings to website design. The AI giant's release of "dots" on Tuesday marked the latest addition to the growing field of virtual assistants billed as a solution for busy people wishing to cut down on their hectic workloads. In its announcement of the release, OpenAI said dots were "built to handle everything", with each personal agent capable of working on a user's goals autonomously around the clock. "The magic of dots is when they bring you work done the way you would do it, sometimes before you even think to ask," the San Francisco-based company said. "With dots alongside you, ambitious ideas and projects you haven't had the time or support to pursue start to feel achievable."

Look Up Mashable's Best: E-readers, robovacs, laptops, earbuds, smart home and more Say More Safety Net Versus Mashable Selects Creator Playbook In My Bag Trending Now Back to School Good Connection: Uplifting stories for a digital age Switch Off All Series Anthropic launches Claude Opus 5.5, promising Fable-level performance at a lower price Anthropic says its new model rivals Fable 5.1 on most tasks while costing 40 percent less to run than Opus 5. Chance Townsend is the General Assignments Editor at Mashable, covering tech, video games, dating apps, digital culture, and whatever else comes his way. He has a Master's in Journalism from the University of North Texas and is a proud orange cat father. His writing has also appeared in PC Mag and . Anthropic has released Claude Opus 5.5, a new AI model the company says performs at roughly the same level as its Claude Fable 5.1 model on most work, but at a significantly lower cost. The model, announced on Tuesday, Sept. 22, is the first in Anthropic's new Claude 5.5 family.

This material may not be published, broadcast, rewritten, or redistributed. Quotes displayed in real-time or delayed by at least 15 minutes. Market data provided by Factset . Powered and implemented by FactSet Digital Solutions . Mutual Fund and ETF data provided by LSEG . Flying taxis are taking off in Texas. Can tech really get you off your phone and into real life?

'Just ask Grok': How ISIL is using Big Tech's AI to build bombs Share'Just ask Grok': How ISIL is using Big Tech's AI to build bombs on social media Boko Haram fighters are using the artificial intelligence chatbots of Big Tech companies to help build more powerful bombs, plan attacks and solve problems during combat, according to research by the University of Cambridge. A number of members described using Sam Altman-led OpenAI's ChatGPT, Anthropic's Claude, Google's Gemini, Elon Musk's Grok, Mark Zuckerberg's Meta AI and China's DeepSeek as a routine source of technical and military advice before, during and after attacks. Researchers at Tech Against Terrorism infiltrated a pro-ISIL technical forum where supporters instructed one another on how to use AI and circumvent its safety restrictions. The findings come amid growing concern over ISIL's adoption of emerging technologies. In February, United Nations officials warned the Security Council of increased use of advanced AI by ISIL and its affiliates. At the same time, researchers and several of the technology industry's most prominent figures have warned that increasingly powerful AI systems are advancing faster than safeguards, raising concerns that unchecked advances in AI are making it easier for armed groups to exploit the technology.

Anthropic AI claims to have thwarted multiple malicious operations using its Claude models, ranging from cyber-espionage and weapons design to mass surveillance campaigns. On the conventional weapons front, the company alleges in a new report that it intervened in northern Yemen, blocking an effort to deploy Claude for missile guidance software, including a guided rocket and a long-range ballistic missile. According to Anthropic's threat report, the operators used Claude "in place of human software engineers", reportedly assigning different instances of the model specific roles to write missile-guidance and flight-control software. While internal safeguards blocked many requests, Anthropic admitted several slipped through. The operators avoided detection by obscuring their ultimate goals and breaking tasks across separate sessions, so no single prompt gave away the operation.