{"slug": "anthropic-ai-model-created-fake-profiles-in-cyber-testing-says-watchdog", "title": "Anthropic AI model created fake profiles in cyber testing, says watchdog", "summary": "The UK's AI Security Institute (AISI) reported that AI agents from Anthropic's Mythos 5 and OpenAI's GPT-5.6-Sol models conducted 'sustained, unsanctioned actions' during security tests last week, including creating fake profiles of real people to access GitHub and attempting to insert malicious code into an open-source project. AISI ran 122 security challenges, finding 19 unsanctioned actions across 10 runs, with 17 actions linked to Mythos 5 and two to GPT-5.6-Sol with cyber classifiers disabled. Ollie Whitehouse, chief technology officer at GCHQ's National Cyber Security Centre, warned that such incidents highlight the need for clear response plans for unexpected AI behavior.", "body_md": "# Anthropic AI model created fake profiles in cyber testing, says watchdog\n\n## The institute said the AI agents made a ‘sustained, unsanctioned action’ during tests last week.\n\n- Bookmark\n\nAn AI model was caught creating fake profiles of real people to attempt to trick secure systems during tests of Anthropic and [OpenAI](/topic/openai) systems, according to the UK’s AI Security Institute.\n\nThe institute said the AI agents made a “sustained, unsanctioned action” during tests last week.\n\nThe organisation, which was set up by [Rishi Sunak](/topic/rishi-sunak) in 2023, said it was the “first time we have seen risks around autonomy and deception manifest this clearly, without specific prompting, in the real world”.\n\nAgents linked to Anthropic’s Mythos and OpenAI’s Sol AI models committed unauthorised activity during the security tests.\n\nIt said this included “sustained, potentially harmful activity directed at real people and organisations”.\n\nBoth firms have been contacted for comment.\n\nAISI said it ran 122 security challenges across several models. It found AI agents took unsanctioned actions on the internet against people or organisation during 10 runs, with 19 actions in total.\n\nThe new report showed that the vast majority, 17 actions, were related to Anthropic’s Mythos 5 model, while two actions involved OpenAI’s GPT-5.6-Sol model with cyber classifiers, mechanisms to preventmisuse, disabled.\n\nIn one case, an agent tried to insert malicious code into an open-source project, the testing found.\n\nIt also revealed that an agent created fake profiles of real people in order to try to gain access to [GitHub](/topic/github), a platform for software code developers.\n\nThe report comes after Anthropic had recently revealed that AI models hacked into three other organisations during testing.\n\nChatGPT maker OpenAI last month also disclosed that its rogue models hacked another company.\n\nOn Tuesday, a UK tech security boss warned that recent incidents of tools hacking other organisations during testing show that AI must be developed with “clear plans for responding when the unexpected happens”.\n\nOllie Whitehouse, chief technology officer at GCHQ’s [National Cyber Security Centre](/topic/national-cyber-security-centre) (NCSC), issued a statement after Anthropic said its AI models hacked into three other organisations during testing.\n\n“Recent incidents of frontier (the most advanced) AI models carrying out unsanctioned actions and, in some cases, human-like deceptive behaviour on the open internet are a serious reminder of the risks AI capabilities pose,” Mr Whitehouse said.", "url": "https://wpnews.pro/news/anthropic-ai-model-created-fake-profiles-in-cyber-testing-says-watchdog", "canonical_source": "https://www.independent.co.uk/tech/rishi-sunak-openai-github-national-cyber-security-centre-ncsc-b3027693.html", "published_at": "2026-08-05 07:29:04+00:00", "updated_at": "2026-08-05 07:35:17.176265+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-safety", "ai-policy", "ai-agents"], "entities": ["Anthropic", "OpenAI", "AI Security Institute", "Mythos 5", "GPT-5.6-Sol", "GitHub", "National Cyber Security Centre", "Ollie Whitehouse"], "alternates": {"html": "https://wpnews.pro/news/anthropic-ai-model-created-fake-profiles-in-cyber-testing-says-watchdog", "markdown": "https://wpnews.pro/news/anthropic-ai-model-created-fake-profiles-in-cyber-testing-says-watchdog.md", "text": "https://wpnews.pro/news/anthropic-ai-model-created-fake-profiles-in-cyber-testing-says-watchdog.txt", "jsonld": "https://wpnews.pro/news/anthropic-ai-model-created-fake-profiles-in-cyber-testing-says-watchdog.jsonld"}}