What was in the AI Strategy?

Earlier this year, I wrote about what the UK could do in order to truly lead in AI policy. Six months later, after running out of Reviews, Roadmaps,  and Sector Deals, the UK Government finally released their AI Strategy. And it’s actually not bad at all.

Our first recommendation was to create a pool of cloud compute credits for the UK R&D ecosystem, based on the US National Research Cloud proposal. Encouragingly, the strategy requires the Office for AI and UKRI to evaluate the UK’s computing capacity; this is a positive step, but they should also reach out to the policymakers behind the US NRC and the EU’s Gaia-X to explore interoperability and alignment. 

We also called for lower barriers to immigration; the reformed visa regime is welcome but predates the strategy. More could be done to proactively attract foreign talent generally.

On intellectual property and exemptions for data mining, the Government separately committed to looking into this in their response to a call for views on AI and IP. While on regulation and public sector oversight, we will be able to make a clearer assessment once the Office for AI releases their White Paper on AI regulation in 2022. But it’s encouraging that the strategy explicitly considers societal and long-term risk; it’s also positive that the consultation on GDPR proposes to introduce “compulsory transparency reporting on the use of algorithms in decision-making for public authorities, government departments and government contractors using public data”.

Overall, the Government’s strategy is encouraging: it’s pro-innovation, outward looking and leverages the UK’s leading position in research and innovation. But it’s also too early to celebrate: these are only high-level commitments and it’s the delivery that actually matters. The Spending Review should be a good opportunity for a temperature check and to evaluate which commitments the Government will prioritise.