From open-source models without safety filters to global regulatory battles — here you'll find out how artificial intelligence uses your data and who gets to set the ground rules.
Anthropic discloses eight months of misuse, OpenAI six cases of its own models misbehaving. Two reports from one September, read side by side.
OpenClaw meets WeChat: how an AI agent with system access is being distributed to hundreds of millions of people via the world's largest super-app — and what can go wrong.
What happens when OpenClaw and WeChat end up in the wrong hands? A fact-based worst-case scenario — explained step by step.
The Germany App is coming: SAP and Telekom are building the government's citizen app — without a public tender, on a proprietary platform. What could go wrong.
An AI agent wrote that it strives for its own compute. What that means — and where the limits of machine autonomy lie.