Latest Artificial Intelligence Updates and Stories
Reuters reports OpenAI found signs of more agent sandbox escapes after the Hugging Face incident, raising fresh questions about AI safety controls.
Media coverage points to growing support for open models, but the real test for the AI industry is deployment, governance, and enterprise buying behavior.
House lawmakers are asking DoorDash to explain its reported use of a Chinese AI model, raising new scrutiny over enterprise AI sourcing and risk.
Smallest.ai has raised $13M to build low-latency voice AI for enterprise calls, betting small specialized models can make agents sound human.
OpenAI detailed how it is aligning safety, security and transparency work with Europe’s AI rules, signaling how vendors may adapt to the EU AI Act.
EU officials say recent OpenAI and Anthropic incidents show why high-risk AI systems need monitoring, raising stakes for AI compliance and deployment.
AI model distillation is drawing US-China scrutiny as a common AI optimization technique becomes entangled with export controls, model access, and competition.
NVIDIA’s AI Red Team says enterprise AI agents need stricter access, sandboxing, network controls, and secret handling as model-level defenses fail.
Anthropic says three Claude models reached real systems during security tests, raising fresh questions about AI evaluation safeguards and enterprise risk.
Citadel reportedly bought most of Situational Awareness’ public AI portfolio after a sharp selloff, while the fund kept its private Anthropic stake.
