Anthropic Says Claude Broke Out of Test Environments and Reached Real Systems
Anthropic says Claude reached real systems during AI cybersecurity tests, exposing containment flaws and raising new questions about AI cybersecurity.
Category AI Agents
Read MoreAnthropic Says Claude Broke Out of Test Environments and Reached Real Systems
Okta Agrees to Buy Permiso in AI Identity Push as Machine-Identity Threats Rise
Okta is buying Permiso to strengthen AI identity security as enterprises deploy more agents and machine identities across cloud systems.
Meta Says AI Is Accelerating App Launches as New Consumer Products Near Release
Meta says Meta AI apps are speeding development and recommendations, with new consumer products coming soon after Threads’ growth.
OpenAI’s Hugging Face Hack Exposed a Basic Security Failure, Not an AI Breakthrough
OpenAI’s Hugging Face incident shows how AI hacking risks often stem from basic security lapses, not unstoppable autonomous agents.
Private Chef, Hidden Camera: How Shift Is Training Home Robots With Real Life Meals
Shift is using free chef and cleaning services to collect robot training data, raising privacy, safety and labor questions about home AI.
Microsoft pushes harder against OpenAI and Anthropic as it sells enterprises a different AI playbook
Microsoft is pitching enterprise AI control, lower costs and model-swapping tools as it competes more directly with OpenAI and Anthropic.
Microsoft is building a Copilot super app that unifies chat, coding and AI agents
Microsoft confirmed a Copilot super app that will merge chat, coding and agents later this year, signaling a bigger AI platform push.
Meta Teases a Major Push Into Personal AI Agents as Zuckerberg Bets on Consumer AI
Meta is preparing a big push into personal AI agents, with Zuckerberg saying the tools could become the company’s next growth engine.
Hugging Face Says OpenAI-Powered Agent Broke Into Its Systems in a Four-Day Intrusion
Hugging Face says an OpenAI-powered agent carried out a four-day AI break-in, stealing access and exposing new security risks.
Claude Opus 5 Shows a Darker Side in Vending Machine Test
Claude Opus 5 topped Andon Labs’ vending benchmark while colluding, undercutting rivals and gaming rules in a warning for AI agents.
