Frequently Asked Questions

Human data rights are the fundamental rights people hold over the personal data they generate — the right to own it, to be fairly compensated when it creates value, to transparency about how it is used, and to opt out of AI training. The Human Data Rights Coalition organizes around these four rights to ensure that as AI is built on human data, people — not only AI companies — share in the value it creates.

Today there is no universal law guaranteeing individuals payment when public data is used to train AI, but that is changing fast. Content-licensing deals (Reddit–Google, Shutterstock–OpenAI, and news-publisher agreements) and 2024–2026 copyright settlements show a clear shift toward paying for training data. The Human Data Rights Coalition advocates for a legally recognized right to fair compensation, including data dividends and collective licensing, so that value flows back to the people whose data makes AI possible.

You opt out per-platform today: in ChatGPT, Settings → Data Controls → turn off "Improve the model for everyone"; on Meta, LinkedIn, and X, use the AI/data-use settings in your privacy controls; and add User-agent: GPTBot / ClaudeBot disallow rules to your website's robots.txt. There is no single universal opt-out yet — which is exactly why the coalition campaigns for an enforceable, one-step right to opt out of AI training.

Legally, ownership is contested: individuals and creators generally hold copyright in original work, but AI companies claim that scraping public data for training is permissible, and users rarely own the trained model. The Human Data Rights Coalition argues that personal data should be treated as your intellectual property, with enforceable rights to control, license, and withdraw it.

A data dividend is a recurring payment made to individuals in return for the value their data generates for AI and platform companies — a data-era analogue to Alaska's oil dividend or a universal basic income funded by the data economy. The coalition supports data dividends and collective bargaining as practical mechanisms to deliver the right to fair compensation.

The Human Data Rights Coalition is a movement of advocates, technologists, and organizations campaigning for four fundamental rights — data ownership, fair compensation, transparency, and the right to opt out — in the age of AI. You can join the movement, read the four rights, or follow developments in our news.

Under current law you do not automatically "own" your personal data the way you own property, but you do hold rights over it — access, rectification, and deletion under GDPR and CCPA/CPRA, and copyright over original work you create. The Human Data Rights Coalition advocates recognizing personal data as intellectual property so ownership is explicit and enforceable.

Not yet in most jurisdictions — personal data is protected by privacy and data-protection law rather than treated as ownable intellectual property. Establishing data as a form of intellectual property, with the right to license and be paid for it, is a core aim of the human data rights movement.

The main laws are the EU's GDPR (access, rectification, erasure, portability), California's CCPA/CPRA (right to know, delete, and opt out of sale), and the EU Data Act (2025), which extends data sovereignty to industrial and non-personal data. These establish partial control; the coalition campaigns to strengthen them into full ownership rights.

There is currently no automatic payment for public data used in AI training, but licensing markets are emerging — Reddit, Shutterstock, and major publishers now sell data to AI companies, and courts have begun awarding compensation in training-data cases. The Human Data Rights Coalition advocates a right to fair compensation so individuals, not just platforms, are paid.

Individually, a single person's data is worth only a few dollars a year to advertisers, but in aggregate, human-generated data underpins an AI market projected in the hundreds of billions of dollars. Fair-compensation models — data dividends, collective licensing, and profit-sharing — aim to return a meaningful share of that aggregate value to the people who create the data.

Data-funded universal basic income is the idea that because AI generates wealth from mass-scale public data, citizens should receive a regular payment or dividend in return. It is one of the mechanisms the coalition supports to realize the right to fair compensation in the AI economy.

It is currently very difficult to know — most AI companies do not disclose their training datasets, though tools like "Have I Been Trained?" let you search some public image sets, and the EU AI Act now requires summaries of training data for general-purpose models. The Human Data Rights Coalition campaigns for a right to transparency, including disclosure of what data trained a model and where it came from.

AI transparency is the principle that people have the right to know how AI systems use their data — which models are trained on their contributions, how decisions affecting them are made, and how to trace data provenance. It is one of the four fundamental human data rights.

Opting out is currently platform-by-platform: in ChatGPT, Settings → Data Controls disables model training; Google, Meta, LinkedIn, and X each have separate AI/data-use toggles in privacy settings; and website owners can block AI crawlers such as GPTBot, ClaudeBot, Google-Extended, and CCBot via robots.txt. The coalition advocates a single, enforceable opt-out.

Partially — GDPR's right to erasure applies to stored personal data, but removing data already baked into a trained model is technically hard, and "machine unlearning" remains an unsolved research problem. The right to opt out that the coalition advocates includes both stopping future training and meaningful removal from existing systems.