Here’s How an AI Slowdown Could Actually Be Enforced
Even if big AI companies agree to a pause, ensuring that nobody tries to sneak ahead could prove tricky.
Recent discussions among AI researchers suggest that a slowdown in AI development may be necessary to ensure the technology remains safe. However, there is still uncertainty about how to effectively enforce such a slowdown. While some propose tighter government regulations and new ways of measuring progress, others suggest more radical measures like tracking devices inside GPUs or even destroying AI chips.
There is growing political and public pressure for a more measured approach to building AI, but experts warn that controlling AI development will require funding and expertise from outside the AI labs themselves. Some ideas, such as giving third-party evaluators greater access to models, have been suggested as potential solutions.
Others believe that more independent and scientifically rigorous inspections may be needed. Experts also point to the need for further research on model evaluations and the alignment of AI models with human values. The role of the US government in restricting AI development is uncertain, but bipartisan support is growing for reigning in the big AI companies.
Trusted Compute is another concept that involves imposing limits on the development of AI through checks on the raw compute required.
Written by urgent.news from Wired Business's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
