ProveIt — AI Content Provenance & Watermarking Compliance Widget for EU Publishers

The Problem

The EU AI Act (Article 50) requires that all AI-generated content — text, images, audio, video — be marked in a machine-readable format, detectable as artificially generated. The labeling requirements become legally binding on August 2, 2026. The EU Commission published a draft Code of Practice (second draft March 3, 2026) specifying a multi-layered approach: metadata embedding, imperceptible watermarking, and visible labeling (a common EU-wide icon or "AI" label).

The problem: most EU publishers, marketing agencies, and e-commerce brands now use AI to generate significant amounts of content (product descriptions, blog posts, social media images, advertising creative). They have zero tooling to:

1. Automatically detect which content in their pipeline is AI-generated
2. Embed C2PA provenance metadata into published content
3. Add machine-readable watermarks to AI-generated images/video
4. Display the required visible label (the EU-mandated "AI" indicator)
5. Maintain audit trails proving compliance to regulators

Big tech (Google SynthID, Adobe CAI) builds these tools for their own ecosystems. There's nothing for the long tail — the WordPress blog, the Shopify store, the agency managing 20 clients.