Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Token Taxes: mitigating AGI's economic risks

Lucas Irwin, Tung-Yu Wu, Fazl Barez

TL;DR

It is argued that the economic risks posed by a post-AGI world can be effectively mitigated by token taxes: usage-based surcharges on model inference applied at the point of sale, which capture value where AI is used rather than where models are hosted.

Abstract

The development of AGI threatens to erode government tax bases, lower living standards, and disempower citizens -- risks that make the 40-year stagnation of wages during the first industrial revolution look mild in comparison. While AI safety research has focused primarily on capability risks, comparatively little work has studied how to mitigate the economic risks of AGI. In this paper, we argue that the economic risks posed by a post-AGI world can be effectively mitigated by token taxes: usage-based surcharges on model inference applied at the point of sale. We situate token taxes within previous proposals for robot taxes and identify two key advantages: they are enforceable through existing compute governance infrastructure, and they capture value where AI is used rather than where models are hosted. For enforcement, we outline a staged audit pipeline -- black-box token verification, norm-based tax rates, and white-box audits. For impact, we highlight the need for agent-based modeling of token taxes' economic effects. Finally, we discuss alternative approaches including FLOP taxes, and how to prevent AI superpowers vetoing such measures.

Token Taxes: mitigating AGI's economic risks

TL;DR

It is argued that the economic risks posed by a post-AGI world can be effectively mitigated by token taxes: usage-based surcharges on model inference applied at the point of sale, which capture value where AI is used rather than where models are hosted.

Abstract

The development of AGI threatens to erode government tax bases, lower living standards, and disempower citizens -- risks that make the 40-year stagnation of wages during the first industrial revolution look mild in comparison. While AI safety research has focused primarily on capability risks, comparatively little work has studied how to mitigate the economic risks of AGI. In this paper, we argue that the economic risks posed by a post-AGI world can be effectively mitigated by token taxes: usage-based surcharges on model inference applied at the point of sale. We situate token taxes within previous proposals for robot taxes and identify two key advantages: they are enforceable through existing compute governance infrastructure, and they capture value where AI is used rather than where models are hosted. For enforcement, we outline a staged audit pipeline -- black-box token verification, norm-based tax rates, and white-box audits. For impact, we highlight the need for agent-based modeling of token taxes' economic effects. Finally, we discuss alternative approaches including FLOP taxes, and how to prevent AI superpowers vetoing such measures.
Paper Structure (9 sections, 1 figure)

This paper contains 9 sections, 1 figure.

Table of Contents

  1. Introduction
  2. Background: The Economic Risks of a post-AGI world
  3. Robot Taxes
  4. The Token Tax
  5. Advantages:
  6. auditing pipeline
  7. Alternative Views
  8. Conclusion
  9. Acknowledgements

Figures (1)

  • Figure 1: Token tax collection procedure. The cloud compute provider acts as an intermediary between the AI model provider and the government. It runs inference on behalf of the AI model provider, collects billed tokens, and determines tax liability through a staged audit pipeline. The provider first attempts (1) a black-box technical token audit. If this fails, it applies (2) a norm-based tax rate derived from average token usage. If the norm-based tax approach fails, it escalates to (3) a white-box audit. Once an audit pathway succeeds, the compute provider confirms the total token tax and reports it to the government.