Securing AI, as it happens

Prompt injection, AI agents and MCP tools, model supply chain, data leaking through assistants, deepfake fraud and the rules arriving to govern all of it. Independent AI security research sits next to what the newsrooms and vendor research teams are reporting.