CONCEPT
Mediators of Individual Data
Lanier and Weyl's proposed institutions — voluntary, member-governed organizations that would aggregate the bargaining power of individual data contributors and negotiate collective terms with AI platforms, analogous to labor unions, professional associations, and music collection societies.
Mediators of Individual Data (MIDs) are Lanier and Weyl's answer to a structural problem that individual rights-based approaches cannot solve: individual data contributors have no leverage against AI platforms, no matter how clearly their rights are articulated. A single novelist whose book was used in training cannot negotiate meaningfully with OpenAI. A single developer whose code was absorbed into Codex cannot extract compensation from GitHub. Individual consent frameworks, individual opt-out mechanisms, and individual litigation are all inadequate to the scale asymmetry. What is required is collective organization — institutions that aggregate the bargaining power of millions of individual contributors into a force capable of negotiating meaningful terms with the platforms that consume their work. MIDs are Lanier's proposal for what those institutions would look like: voluntary, member-governed, representing contributor interests, negotiating royalty rates and usage terms, distributing payments, and enforcing compliance. The historical precedents are well-established. The specific application to AI training data is not yet built.
