The UK AI Council, a non-statutory expert committee of independent advisers set up in 2019 published its AI Roadmap today. I’ve long been interested – and occasionally frustrated – by the issue of so-called AI nationalism. But there are undeniably hotspots of AI and London was certainly the pre-eminent one pre-Brexit in, as I noted almost three years ago in this post. It’s too early less than a week post-Brexit to determine what happens in that regard, but the AI Council attempts to lay out a roadmap for the government.
It divides its recommendations into four areas:
- Research, Development & Innovation
- Skills and Diversity
- Data, Infrastructure and Public Trust
- National, Cross-sector Adoption
A brief summary of each one:
- Research, Development & Innovation – talent development & retention from around the world (ahem, Brexit) is crucial and the government should ensure The Alan Turing Institute (the UK’s national institute for data science and AI) draws talent from across the UK, not just the usual spots (e.g. London, Cambridge, Oxford, Edinburgh); encourage moonshots, through it describes explainable AI as one of those, which I’m a bit skeptical if that still qualifies as a moonshot, but the report detail lists “those involving machine learning from small data sets, and machine learning as part of wholly new types of neural interface technologies”
- Skills and Diversity – 10-year high-level skill-building program, including an Online Academy for AI and “AI-related Level 7 apprenticeships, Masters programmes, Centres for Doctoral Training and novel forms of PhDs,” as well as al round data literacy programmes. In terms of diversity, the report recommends benchmarking diversity as is done already in the National Cyber Security Centre.
- Data, Infrastructure and Public Trust – consolidate & accelerate infrastructure; make the UK a leading authority on data governance & ensure public trust through public scrutiny. The last of those isn’t something the UK ‘leads the world’ on, imho. But the report also recommends “strengthen[ing] existing partnerships with other like-minded countries, such as France and Canada.”
- National, Cross-sector Adoption – support local initiatives support startups, enable robust public sector investments, use AI to meet challenges of Net Zero carbon emissions (one of those moonshots); in defence make sure AI helps keep the country safe and – building on the work of NHSX – keep it as healthy as possible. This last part, about industry-specific has some of the most cogent recommendations, I feel. The report also says “the UK currently lacks a prominent cohort
of professionals who understand the complex sales and leadership skills needed to build, articulate and sell AI tools,” which is a bit unfair to some of my friends and former colleagues in the industry because I think that could be said about any nation on earth, right now. But it will need addressing, as will product management and other skills.
At one point the report says one of the goals of the AI Council is “for everyone to be able to live confidently with AI.” Amen to that.