The importance of AI Compute for National AI strategies
20, March, 2023
A recent report from the OECD Directorate for Science, Technology and Innovation (STI) questions whether countries have enough compute capacity to achieve national AI strategies. The authors have examined the compute capacity required by AI and posit that it is time for countries to consider this when forming their AI strategies. Indeed the lack of this is currently a policy blind spot.
The report is a timely reminder of the importance to plan for the development of AI and the necessary infrastructure that will be needed to support this. It is also essential to develop this infrastructure in line with the EU Green Deal and digital transition to ensure that EU countries both take full advantages of the potential afforded by AI in a sustainable manner. Some key takeaways from the report include:
1. Artificial intelligence (AI) can provides economic opportunities for productivity, growth, and resilience.
2. It is clear that some countries have developed their AI strategies without fully assessing whether they have sufficient domestic AI compute infrastructure and software to realise their goals.
3. The demand for AI compute is growing dramatically and will continue to do so.
4. AI compute divides are expected to emerge or deepen in the coming years.
In the report the OECD proposes a "AI compute" blueprint for taking stock of the indicators, datasets, and proxies countries will need to measure national AI compute capacity. It also identifies challenges to measuring and benchmarking national AI compute capacity across countries.
You can download and read the full report here: https://oecd.ai/en/compute-report