Provenance Pass
A registry where writers, photographers, illustrators and developers can register their portfolio and set licensing terms for AI training — companies that want to train or fine-tune can query the registry, get a clean license, and the creator gets paid automatically.
Independent creators (writers, photographers, illustrators, indie devs) worried about AI using their work without consent or pay
- One-click registration of a creator's portfolio (blog, GitHub, Flickr, portfolio site) with cryptographic proof of authorship and timestamp
- Tiered license catalog (free with attribution, paid commercial, full opt-out) that AI companies can browse and license via API
- Micropayment payouts every time a registered work is included in a verified training run, with a public dashboard for the creator
- Opt-out registry that AI labs commit to honor before any new training, with an audit log of who checked what and when
Andrei Merlescu's essay captures a turning-point sentiment: creators are asking why they should keep publishing if their work trains the systems that displace them — a registry that turns open contributions into licensed, paid data gives creators a concrete reason to keep publishing and gives AI labs a clean way to source.
Strong, measurable demand: 11,000+ creatives signed an open letter against unauthorized AI scraping, Spawning's opt-out registry was respected by Stability AI for SD3 training, and a recent investigation revealed Meta/Google/Nvidia scraped 15M+ YouTube videos without consent — creators are actively seeking consent + compensation mechanisms.More than 11,000 creatives condemn unauthorized use of content for AI ↗Have I Been Trained is Back! - by Laura Exline - Substack ↗
Multiple direct or near-direct competitors already exist: Credtent (rights-cleared licensing marketplace), DECORAIT (decentralized opt-in/out registry with rewards), PersonaShare (creator AI-licensing marketplace), and Spawning (Do Not Train registry respected by Stability AI) — this is an actively contested space, not wide open.Credtent: Neutral Licensing Infrastructure for the AI Economy ↗DECORAIT - DECentralized Opt-in/out Registry for AI Training ↗AI Likeness Licensing Marketplace | PersonaShare ↗
Real willingness to pay on the AI lab side: Shutterstock generated $104M in a year licensing assets to AI developers, and the same article describes the Contributor Fund paying creators — but a standalone indie-creator registry hasn't proven unit economics, and creators face friction/low expected payouts per work.Shutterstock Made $104 Million Licensing Assets to AI Devs Last Year ↗Shutterstock Data Licensing and the Contributor Fund ↗
Tailwinds are structural: regulators in the EU (TDM opt-out) and Canada are codifying consent/compensation rules, lawsuits are mounting, and AI training data demand will only grow as model builders seek legally clean corpora — the need persists for years.Consultation on Copyright in the Age of Generative Artificial Intelligence - Canada ISED ↗Investigation reveals unauthorized data scraping from YouTube for AI training ↗
Hard to build credibly: requires identity/KYC for creators, legal contract templates per content type, payment rails, API integrations with model-training pipelines, and adoption by AI labs — competing against funded players (Spawning, Credtent) and incumbent stock libraries with existing catalogs and trust.DECORAIT - DECentralized Opt-in/out Registry for AI Training ↗