Your Electricity Bill Could Help Spot Secret AI Training
New research says power use might reveal hidden AI, but it's not foolproof.
Deep Dive
arXivLabs is a framework that lets collaborators develop and share new arXiv features directly on the arXiv website. Both the individuals and the organizations that work with arXivLabs have embraced and accepted arXiv's values of openness, community, excellence, and user data privacy.
arXiv says it is committed to those values and only works with partners who adhere to them. If you have an idea for a project that would add value for arXiv's community, the page points you to learn more about arXivLabs.
Key Points
- Monitoring electricity use is not a reliable way to detect secret AI training because many other activities also use lots of power.
- Power monitoring can be easily evaded by training AI in small bursts or in different locations.
- Effective AI oversight will need a combination of methods, not just one, and international cooperation.
Why It Matters
Without reliable detection, we can't enforce rules to prevent AI misuse, affecting safety and trust.