The token economy requires collective data governance

The token economy requires collective data governance. Millions of internet users generate the raw material that AI models are trained on, but they rarely share in the wealth created. For over two decades, digital platforms have profited from our attention and social interactions, yet now, AI models are extracting value from our collective knowledge.

Every piece of content, from articles to videos and comments, can be turned into training data, or "tokens," ultimately leading to value creation. This new phase of digital capitalism is known as the "token economy."

Unlike the attention economy, which capitalized on capturing users' attention, the token economy turns human knowledge and creativity into data for AI training. Since this token economy relies on content produced by millions of users, the rules governing AI training data cannot be solely controlled by a handful of tech firms. The rules must be governed as a digital common resource and managed through collective institutions.

Collective-management organizations, which have been around for over a century, could serve as a model. These organizations negotiate licenses on behalf of millions of authors and artists, collect royalties, and redistribute revenues according to transparent rules. Instead of each user negotiating separate agreements with tech companies, these organizations could represent internet users, creators, and rights holders.

They would negotiate access to content fed to large language models, determine the value's distribution, and ensure tech companies respect users' choices.

However, the role of these organizations could extend beyond compensation. They could also determine the conditions under which AI developers access data, allowing members to shape the values embedded in these systems and ensure they serve the public interest. By becoming stakeholders in AI governance, users would no longer be passive sources of data but active participants in the decision-making process.