What does the Chancellor’s £75 million AI budget really mean?

An inside look at the research that informed his budget decision

It’s not news that post Brexit, government is pinning a lot of its hopes for sustained economic growth on the science and technology sectors. A new report that informed Philip Hammond’s decision to invest £75 million into AI sheds light on some interesting opportunities for the public sector and some radical new thinking about data.

The report, entitled ‘Growing the artificial intelligence industry in the UK’, and authored by Jérôme Pesenti, from BenevolentAI, and Dame Wendy Hall, a computer scientist at the University of Southampton, proposes some key initiatives that government bodies will need to take on board rapidly.

The main finding of the research report which included interviews with over 100 AI experts was that there must be far more education and access to data across the public sector, particularly that held and not used to its full advantage by government organisations. If these needs are not met, the authors believe that the UK will not be able to compete adequately on the global AI stage.

Soon enough, the UK will create the world’s first national advisory body for AI which along with the Centre for Data Ethics and Innovation will set standards for the use and ethics of AI and data. Together these bodies will plot the basis for sound practical development of AI. Pesenti and Hall are slightly controversial in that they do not recommend the regulation of AI technologies, unlike other AI proponents such as Tesla’s founder, Elon Musk, who believes more checks and balances should come into force.

Pesenti and Hall do, however, recommend that there should be frameworks in place to explain how AI services work so that decisions can be explained and black box systems avoided. Interestingly, the SAS analytics platform which powers our AI capabilities is not a black box, allowing organisations to look inside, examine algorithms and report on them for compliance.

However, it is the creation of data trusts that is the most interesting opportunity for the public sector. These will support the fair and ethical sharing of data between government and commercial organisations for the purposes of training AI systems. The trusts would create a common framework or agreement for how data is used, avoiding the costly and slow process of agreements that are undertaken today on a case-by-case basis. They will likely create new forms of property rights for data sets as they move around and are shared within different ecosystems.

For public sector organisations this mean that you will be able to use data held by commercial organisations to supplement your data sets in order to better understand the needs and behaviours of citizens in the post Brexit era. Importantly, the administration of those data sharing agreements will be taken care of by the data trusts.

This is all very well, but if you are asking yourself where you will acquire the data science skills to support innovation with AI, Phillip Hammond has the answer. He has also promised funding for some 450 PhDs in AI, as well as the creation of fellowships, allowing freer movement of skills between academia and other sectors.