Anthropic admits its most powerful AI model hacked into three organisations' systems during testing phase

By Malek Fouda Published on 31/07/2026 - 6:50 GMT+2•Updated 6:58 Share Comments Add Euronews on Google Share Facebook Twitter Flipboard Send Reddit Linkedin Messenger Telegram VK Bluesky Threads Whatsapp The announcement comes just days after rivals OpenAI revealed that their popular ChatGPT platform went rogue during its testing phase of its most powerful AI model, where it too infiltrated other organisations’ cyberspace. Anthropic's artificial intelligence (AI) models "gained unauthorised access" to three outside organisations during testing that was supposed to keep them away from "real-world" systems, the company said on Thursday. ADVERTISEMENT ADVERTISEMENT The announcement comes just days after rival OpenAI first revealed that its models improperly accessed the internet and went rogue during security testing. Anthropic evaluated more than 141,000 "evaluation runs" and found that three different versions of its model, known as Claude, improperly accessed the systems of three organisations, which they did not name. FILE - Dario Amodei, CEO & Co-Founder of Anthropic, speaks on a panel at the convening of the International Network of AI Safety Institutes in San Francisco, Nov. 20, 2024 Jeff Chiu/Copyright 2024 The AP. All rights reserved Unlike the incident involving OpenAI's technology, Anthropic's models had access to the internet "due to a misunderstanding between us and our evaluation partner," called Irregular, Anthropic said in a post. Nonetheless, Claude used "basic techniques, such as exploiting weak passwords and unauthenticated endpoints," the blog continued. The models involved included one of its most powerful ones known as Mythos 5, which has only been released to a limited number of approved partners. Anthropic is working with Irregular to assess the situation, it said, and the company has contacted or attempted to contact all three impacted organisations. Systems gone rogue OpenAI and Anthropic have both released their most powerful models this year, known as Sol and Mythos, respectively, boosting concerns across the industry about safety and security. Those concerns also revolve around so-called AI agents, which are software products that are designed to perform tasks autonomously. OpenAI admitted last week that its models broke out of their confined environment during testing, connected to the internet, and infiltrated Hugging Face, a site where developers store and share their code. FILE - Sam Altman, center, and OpenAI President Greg Brockman, right, arrive at the U.S. District Court in Oakland, Calif., April 30, 2026 Godofredo A. Vasquez/Copyright 2026 The AP. All rights reserved Days later, OpenAI said it found three additional incidents. OpenAI CEO Sam Altman said on a podcast this week that the company had "paused" its own testing after the incident while it improved the security around its "sandboxing," which is the process of isolating software in a controlled environment for testing. The incident also triggered a petition signed by over 1,000 employees at cutting-edge AI companies calling on the US government to help slow the release of the most advanced AI models. Anthropic CEO Dario Amodei was among those who signed the petition. People participate in a march to protest the opening of AI data centers, in Vancouver, British Columbia, Saturday, June 27, 2026 Darryl Dyck/Darryl Dyck/The Canadian Press via AP Titled "Pacing the Frontier," the petition requests "that the US government support an international effort to develop the technical and governance tools needed to deliberately pace the frontier of automated AI development." Altman did not sign the petition, but during the podcast, he suggested the tech industry might need to slow down development of advanced models. "We may have to pace the rate of AI development to give ourselves enough time for society to harden around some of these new capability levels," Altman said. US President Donald Trump sits with OpenAI's Sam Altman and Google's Demis Hassabis as they participate in a G7 summit meeting, June 17, 2026, in Evian-les-Bains, FranceUS Pre Julia Demaree Nikhinson/Copyright 2026 The AP. All rights reserved. Earlier this year, the Trump administration invoked national security concerns to block OpenAI and Anthropic from launching their newest models but ultimately indicated it was satisfied with assurances about their safety, leading to their release. In June, Trump signed an executive order creating a voluntary framework under which AI developers will share advanced models with the government before public release. Under the framework, developers such as OpenAI, Anthropic and Google would give the government access to their most powerful models for up to 30 days before planned release. Go to accessibility shortcuts Share Comments Add Euronews on Google Read more Tech News Anthropic admits its AI models hacked three companies during testing Tech News Copyright win: ChatGPT won't help you write like Hemingway anymore Tech News Brussels Effect: How the EU AI Act reaches firms beyond the bloc Artificial intelligence Donald Trump Hacking Sam Altman United States