As artificial intelligence (AI) continues to transform various industries, a comparison gains traction: considering AI as a resource akin to oil or minerals. This comparison illuminates the idea that AI capitalizes on human knowledge, similar to how industries exploit natural resources. This analogy underscores the concept of implementing a tax policy to secure the benefits of this resource for the public good, similar to models in place for natural resource extraction.
The foundation of AI models—data derived from human works—positions them as a resource owned collectively by society. Much like severance taxes on oil extraction, which benefit public funds, it seems equitable for companies profiting from AI to contribute proportionately. These taxes would ideally fund a sovereign wealth fund, mirroring the approach that states like Alaska use for natural resources. Information on state oil and gas severance taxes can be found here.
The principle is not unprecedented. Alaska’s Permanent Fund is a prime example, instituted to preserve wealth derived from oil taxes for future state residents. Established in 1976, the fund currently offers annual dividends to Alaskan residents, with the 2024 dividend set at $1,702. Additional details about the Alaska Permanent Fund can be accessed here. Similarly, Norway’s Government Pension Fund, financed through oil and gas taxes, provides long-term economic stability and holds nearly $1.8 trillion in assets, as evidenced here.
The AI ecosystem, while advancing societal benefits, centralizes wealth among a few corporations, presenting a stark contrast to the public nature of the data used. Implementing a tax on AI models, particularly on their parameters, could ensure equitable distribution of the generated wealth. Parameters are the core elements in language learning models that encapsulate extracted information. Initial tax rates could be minimal to avoid hampering innovation. For further insights on AI taxation methods, see the discussion here.
Directing the resulting revenue into a U.S. sovereign wealth fund could help maintain the public value derived from AI advancements. Such a model, although unlikely to match Norway’s fund in size, could still offer substantial future benefits, including subsidies for essential services and investment in infrastructure. As AI technology progresses, instituting these tax policies sooner rather than later ensures the benefits of this collective resource are not monopolized by private entities.
To delve deeper into Andrew Leahey’s arguments and the broader context of AI taxation, read the full article here.