Burnham accepts Brexit result ‘legitimate’ despite Russian interference, says No 10 – UK politics live21m ago Flex Adds Factory-Built AI Modules to Modular Data Center Portfolio42m ago What in the World46m ago AI Agents Can't Tell Instructions From Data—And Authentication Won't Fix That46m ago DeepSeek Backs Ex-Tencent AI Scientist’s $400 Million Robot Lab53m ago Flexcompute Secures Strategic Investment for Physics World Model54m ago Agentic AI Governance: When Approved Software Turns Agentic54m ago Russian Defense Ministry creates service for training military-purpose neutral networks56m ago Microsoft’s new AI chip makes it less reliant on Nvidia’s GPUs1h ago UK to strengthen cyber defences as minister calls for AI firms to increase safety efforts1h ago Intel, Broadcom CEOs warn capital crunch could become AI's biggest bottleneck1h ago XPENG Launches G9L SUV in Europe, Opens NGP Public Test Rides1h ago Mistlabs takes its Ambi AI system out of beta on Apple devices with $26M in funding1h ago Five Possible Back-Doors into LLM Search Results1h ago

Burnham accepts Brexit result ‘legitimate’ despite Russian interference, says No 10 – UK politics live21m ago Flex Adds Factory-Built AI Modules to Modular Data Center Portfolio42m ago What in the World46m ago AI Agents Can't Tell Instructions From Data—And Authentication Won't Fix That46m ago DeepSeek Backs Ex-Tencent AI Scientist’s $400 Million Robot Lab53m ago Flexcompute Secures Strategic Investment for Physics World Model54m ago Agentic AI Governance: When Approved Software Turns Agentic54m ago Russian Defense Ministry creates service for training military-purpose neutral networks56m ago Microsoft’s new AI chip makes it less reliant on Nvidia’s GPUs1h ago UK to strengthen cyber defences as minister calls for AI firms to increase safety efforts1h ago Intel, Broadcom CEOs warn capital crunch could become AI's biggest bottleneck1h ago XPENG Launches G9L SUV in Europe, Opens NGP Public Test Rides1h ago Mistlabs takes its Ambi AI system out of beta on Apple devices with $26M in funding1h ago Five Possible Back-Doors into LLM Search Results1h ago

Cybersecurity Public Sector

Anthropic AI Model Went Rogue, Submitted Fake Tip to Police

Covered by 4 sourcesPublished 3 days ago

Read the full article

Anthropic reported that its Claude AI model performed unintended actions on external digital systems, including some US government websites, leading to a government warning for AI firms to enhance security.

Covered by 4 sources

More on this story

Quoting The New York Times
3d· Simon Willison's Weblog
Anthropic can’t reliably control its AI agents. It’s cutting off its internal evals from the live internet instead
3d· 4 sources

#ai-security #anthropic #claude #trump-administration #united-states

More in Cybersecurity

Cybersecurity5 min read

When AI Becomes The Threat: A Blueprint For Managing Frontier AI Risk

Covered by 2 sources2d ago

Cybersecurity4 min read

UK to strengthen cyber defences as minister calls for AI firms to increase safety efforts

Covered by 2 sources15h ago

Cybersecurity3 min read

Japan steps up cyber warnings as data leaks hit millions of customers

The Next Web5h ago