compute governance
In plain wordsUsing control over the computer chips and data centers that AI depends on — who can buy them, where they go, how they are used — as a way to oversee powerful AI.
The use of computing power (computecompute, n.: The raw computing power used to train and run AI, usually measured by how many specialized chips are used and for how long. With data and algorithms, it is a main driver of AI progress.) — AI chips, the data centers that house them and the cloud services that rent them — as a point of leverage for AI policy. In the framing of Girish Sastry, Lennart Heim and colleagues (2024), compute is an unusually workable lever because it is 'detectable, excludable, and quantifiable' and comes from a concentrated supply chain: big training runs need big, power-hungry facilities that are hard to hide; chips are physical goods that can be counted and withheld; and a handful of firms such as NvidiaNvidia, prop. n.: An American chip company whose graphics processors became the main hardware for training and running AI, making it one of the most valuable companies in the world., TSMCTSMC, prop. n.: The Taiwanese company that manufactures most of the world's advanced computer chips, including nearly all of the specialized chips used to train and run AI. and the Dutch lithography maker ASML control key stages of production. Tools include export controls on AI chipsexport controls on AI chips, n.: Government rules that limit which countries and companies may buy the most advanced AI chips and the machines used to make them. The United States uses them mainly to slow China's progress in AI., reporting duties keyed to compute thresholdscompute threshold, n.: A legal cutoff based on how much computation was used to train an AI model. Models trained with more than that amount face extra rules, on the theory that bigger training runs produce more capable models., know-your-customerknow-your-customer for compute, n.: Rules that would make cloud companies check who their customers are before renting them enough computing power to train a large AI model, much as banks must verify who opens an account. rules for cloud providers, chip registries, on-chip governance mechanismson-chip governance mechanisms, n.: Features built into AI chips that could check or enforce rules about how they are used — for example, proving which country a chip is in, or refusing to run without a valid license. and public compute for researchers. Critics note that algorithmic efficiency gains, distillationknowledge distillation, n.: Training a smaller or cheaper 'student' AI to copy the behavior of a larger 'teacher' AI, so the student keeps much of the teacher's skill while being faster to run. and growing inference-time computetest-time compute, n.: The computing effort an AI model spends while answering a question, as opposed to while being trained. Letting a model think longer or try many answers can make it markedly smarter. weaken training compute as a proxy for risk, and that controls invite smuggling and evasion.
Compute governance works only as long as frontier training still needs tens of thousands of top-end chips.
Origin
Discussed by researchers at the Centre for the Governance of AI (GovAI) and elsewhere from about 2021. Yonadav Shavit's 2023 proposal for verifying training runs through compute monitoring and the multi-author paper 'Computing Power and the Governance of Artificial Intelligence' (Girish Sastry, Lennart Heim et al., 2024) gave the field a systematic form; US chip export controls (2022) and Executive Order 14110Executive Order 14110, prop. n.: A sweeping 2023 order by US President Joe Biden that told federal agencies how to manage AI risks and required makers of the largest AI models to report their safety tests to the government. It was revoked in 2025. (2023) were early applications.
Related
- compute thresholdcompute threshold, n.: A legal cutoff based on how much computation was used to train an AI model. Models trained with more than that amount face extra rules, on the theory that bigger training runs produce more capable models.
- export controls on AI chipsexport controls on AI chips, n.: Government rules that limit which countries and companies may buy the most advanced AI chips and the machines used to make them. The United States uses them mainly to slow China's progress in AI.
- know-your-customer for computeknow-your-customer for compute, n.: Rules that would make cloud companies check who their customers are before renting them enough computing power to train a large AI model, much as banks must verify who opens an account.
- on-chip governance mechanismson-chip governance mechanisms, n.: Features built into AI chips that could check or enforce rules about how they are used — for example, proving which country a chip is in, or refusing to run without a valid license.
- AI Diffusion RuleAI Diffusion Rule, prop. n.: A 2025 US rule, cancelled before it took effect, that would have sorted the world's countries into three groups with different limits on how many advanced AI chips they could buy.
- training computetraining compute, n.: The total number of arithmetic operations performed to train a model. It is the standard yardstick for the size of an AI training effort, and some laws use it to decide which models face extra rules.
Sources
- Sastry, G., Heim, L., Belfield, H., Anderljung, M., Brundage, M., Hazell, J., O'Keefe, C., Hadfield, G. K. et al. (2024). Computing Power and the Governance of Artificial Intelligence. arXiv:2402.08797.
- Shavit, Y. (2023). What Does It Take to Catch a Chinchilla? Verifying Rules on Large-Scale Neural Network Training via Compute Monitoring. arXiv:2303.11341.
Mentioned in
AI governanceAI governance, n.: The rules, institutions and habits that shape how AI is built and used — from laws and treaties to technical standards and a company's own review processes., computecompute, n.: The raw computing power used to train and run AI, usually measured by how many specialized chips are used and for how long. With data and algorithms, it is a main driver of AI progress., export controls on AI chipsexport controls on AI chips, n.: Government rules that limit which countries and companies may buy the most advanced AI chips and the machines used to make them. The United States uses them mainly to slow China's progress in AI., CHIPS and Science ActCHIPS and Science Act, prop. n.: A 2022 US law that put tens of billions of dollars into building computer-chip factories and chip research in the United States, to reduce dependence on chips made in Asia., compute thresholdcompute threshold, n.: A legal cutoff based on how much computation was used to train an AI model. Models trained with more than that amount face extra rules, on the theory that bigger training runs produce more capable models., concentration of powerconcentration of power, n.: The worry that AI will put enormous economic, political or military power in the hands of a few companies, governments or individuals, beyond the reach of ordinary checks and balances., Epoch AIEpoch AI, prop. n.: A research nonprofit that gathers data on how fast AI is advancing — the computing power, data and money behind it — and publishes forecasts and difficult tests for AI., AI Diffusion RuleAI Diffusion Rule, prop. n.: A 2025 US rule, cancelled before it took effect, that would have sorted the world's countries into three groups with different limits on how many advanced AI chips they could buy., Center for Security and Emerging TechnologyCenter for Security and Emerging Technology, prop. n.: A Georgetown University think tank in Washington, D.C., whose researchers advise the US government on AI and national security, paying special attention to China and to the chip supply chain., Centre for the Governance of AICentre for the Governance of AI, prop. n.: A research organization, based in Britain, that studies how governments and companies should handle the dangers of advanced AI and the large economic and political changes it could cause., CERN for AICERN for AI, n.: A proposed publicly funded, international AI laboratory modeled on CERN, the European particle-physics center, where countries would pool money, computers and scientists to do AI research together., decentralized trainingdecentralized training, n.: Training an AI model on computers owned by many different people or companies around the world, linked over the ordinary internet rather than housed together in one data center., GPU poorGPU poor, adj.: Slang for researchers, startups or countries without access to the huge numbers of expensive chips needed to build the most advanced AI, as opposed to the GPU-rich big labs., IAEA for AIIAEA for AI, phr.: The idea of an international agency for the most advanced AI, modeled on the UN's nuclear watchdog, which inspects countries' nuclear programs to keep them safe and peaceful., know-your-customer for computeknow-your-customer for compute, n.: Rules that would make cloud companies check who their customers are before renting them enough computing power to train a large AI model, much as banks must verify who opens an account., MolochMoloch, prop. n.: A name, borrowed from an ancient god of sacrifice, for situations where competition forces everyone to act against their shared interests — as when AI companies feel pushed to race ahead despite the risks., Mutual Assured AI MalfunctionMutual Assured AI Malfunction, n.: A proposed deterrence strategy in which any country racing for overwhelming AI dominance would expect rivals to sabotage its AI projects, so none tries, much as with nuclear deterrence., National AI Research ResourceNational AI Research Resource, prop. n.: A US program to give university researchers, students and small companies shared access to the costly computers, data and AI models that are otherwise mostly available to big technology firms., on-chip governance mechanismson-chip governance mechanisms, n.: Features built into AI chips that could check or enforce rules about how they are used — for example, proving which country a chip is in, or refusing to run without a valid license., remote attestationremote attestation, n.: A way for a computer to prove to someone far away exactly what software it is running, using a signature built into its hardware, so they can decide whether to trust it.