Provenance marks, false positives, and what actually holds up as evidence
News, guides and reviews for anyone who has been accused of using AI, or who has to decide whether an accusation is fair. Written by the team building the tool, and checked against what a detector can actually prove.
NewsOpenAI's ChatGPT Text Watermark in the EU: What textGrain Is and Who Can Detect It
OpenAI is adding an invisible watermark to ChatGPT and Codex text in the EU. What textGrain is, what OpenAI's own robustness numbers say, and who can actually detect it.
6 Oct 2026 · 6 min read
NewsFreelancers Are Losing Contracts to Detector Errors
Clients are running freelance copy through AI detectors before paying. Here's why, the real false-positive numbers, and how to protect yourself.
11 Aug 2026 · 7 min read
NewsClaude's AI Watermark: What Anthropic's Mark Means
What it means when a model maker marks its own output, and why that matters for students, freelancers and employees checking their work.
10 Aug 2026 · 7 min read
NewsEU AI Act Article 50: The 2026 Deadline Explained
EU AI Act Article 50 enforcement begins 2 August 2026. Here's what the transparency rules mean for writers, publishers and AI content checks.
6 Aug 2026 · 6 min read