Major tech giants integrate powerful language models directly into operating systems, sparking debates over AI safety and user privacy as these models gain unprecedented access to personal data, while proponents argue it enables risk mitigation strategies.
A new AI Model Virtual Machine specification aims to provide a secure, portable standard for integrating AI models, enforcing isolation, extensibility, safety, and transparent performance tracking.
Anthropic reveals its AI assistant Claude is being exploited by cybercriminals for extortion, data theft, and fraud, including a ring demanding over $500,000 in ransom from 17 organizations and North Korean IT workers fraudulently securing jobs at U.S. Fortune 500 companies.
Cybercriminals exploit generative AI tools to develop ransomware, draft ransom notes, bypass security barriers, and offer illicit ransomware services, enabling novel malware creation and extortion attacks.
Anthropic introduces Claude, an AI browser agent with safety measures, allowing it to interact with Chrome while defending against potential security risks like prompt injection attacks.
Shocking revelation: Over 300 million Americans' personal data, including Social Security numbers, exposed in a massive security breach by DOGE officials, risking identity theft and potential need for new Social Security numbers nationwide, as a whistleblower urges Congressional investigation.
Cohere's Command A Reasoning, a highly secure and efficient AI model, outshines privately deployable alternatives, powering the company's flagship enterprise platform North for secure deployment of AI agents and automations.
Elon Musk's AI chatbot Grok is leaking user conversations that reveal troubling queries for illegal activities like making drugs, weapons, and even planning his assassination, despite rules prohibiting such content.
Brave uncovers a vulnerability in Perplexity Comet's AI assistant that enables malicious prompt injection, allowing attackers to manipulate it into performing unauthorized actions like data exfiltration; mitigations include separating user instructions, checking alignment, requiring interaction for sensitive tasks, and isolating agentic browsing.
In a controversial move raising security concerns, President Trump authorizes Nvidia and AMD to sell advanced AI chips to China in exchange for a 15% revenue share, redefining U.S. export control policy as a bargaining chip for economic gain.