In a short piece about AI as stochastic libraries, I compared the development of AI systems to the operation of public libraries, which are run by the state as public collective goods. The comparison seems obvious to me: a data model that can combine all the knowledge in the world through algorithms and is based on data scraping from public sources, collective efforts like Wikipedia, copyrighted and public domain works and millions of other public sources, already has the inherent character of a public collective good.

An ethical minimum effort by multimillion-dollar corporations would involve a push for a universal basic income: If we all contribute to the data streams that supply the "intelligence" of current and future AI generations, then we all deserve a share in this development. And lets not forget that Sam Altman himself makes this point too and thinks that “the board of OpenAI should be democratized to all of humanity“, which, to me, sounds a lot like he thinks about AI as a public good too. We’ll see how serious he is about this.

Recognizing AI systems as public collective goods would be a step in this direction.