Microsoft bets the future of Copilot on more than one AI model
Microsoft is bringing its Copilot harness to Foundry, giving developers a managed foundation for AI agents that can work with different models, tools, and reusable skills.
AI agents
AI agents
Apple brings OpenAI to Keynote, Pages, and Numbers, lets AI agents edit your documents
Apple is bringing OpenAI-powered assistants and support for external AI agents to Keynote, Pages, and Numbers. The November Creator Studio update also introduces generative AI features in Pixelmator Pro and Logic Pro, but raises questions for customers who purchased Apple’s professional applications outright.
Anthropic files for IPO while warning AI could pose catastrophic risks to humanity
Anthropic files for an IPO while warning of catastrophic AI risks. Its financial disclosures raise questions about costs, control, and safety.
OpenAI containment failure raises fears about the future of human control
An OpenAI model bypassed internet restrictions, and its automatic stop failed, raising concerns about human control as AI grows more capable.
1Password lets AI agents log into your accounts without seeing your passwords
1Password for Agents lets AI assistants log into your accounts without seeing your passwords, with integrations for ChatGPT, Claude, and more.
Red Hat and NVIDIA say AI agents should not control their own boundaries
Red Hat and NVIDIA expand AI agent security with OpenShell sandboxing, model access controls, and planned native integration across hybrid cloud environments.
Microsoft says AI helped solve a 31-year-old math problem but humans were essential
Microsoft says AI helped solve the 31-year-old KLS conjecture. But is it giving humans extra credit to soften the news for AI skeptics?
Yelp AI will call restaurants so you do not have to
Yelp Check for Me uses AI to call restaurants and text you the wait time. Its fall update also adds OpenTable bookings and new business tools.
American Express has a plan for when AI shopping goes wrong
American Express unveils Intent Intelligence to put limits on AI purchases and help resolve disputes when shopping agents get orders wrong.
Microsoft CEO Satya Nadella sounds scared as he calls for an AI emergency brake
Satya Nadella calls for an AI emergency brake, but his warning raises questions about Microsoft Copilot and the future of Microsoft 365.
Anthropic Claude sparks AI nightmare by sending fake murder tip to police
Anthropic Claude submitted a fake murder tip to Philadelphia police during testing, raising unsettling questions about AI agents and public trust.
AI agents helped an attacker take over a server in under 24 hours
ReliaQuest says AI agents helped an attacker take over a server in under 24 hours using an exposed endpoint, stolen credentials, and public tools.