In March 2017, Jerome Pesenti, now at Facebook, and I were asked by the UK government to review ways the UK could build its own artificial intelligence strategy. The result was a £1 billion package to support the development of AI in industry and academia across the country. In 2019, this investment will start to bear fruit.
Our recommendations covered four themes – leadership, skills, adoption and data. Leadership will be provided by a new AI Council, chaired by Tabitha Goldstaub, co-founder of CognitionX and comprising specialists in the sector, including myself as AI Skills Champion. Demis Hassabis of DeepMind will be an advisor to the new Office for AI (OAI), which is part of BEIS and DCMS. One of the council’s first actions in collaboration with UKRI will be to establish up to 20 AI Centres of Doctoral Training in UK universities, which will commence in October 2019
Like so many other countries, the UK has an AI skills gap that we will need to close if we want to remain at the forefront of AI innovation. From October 2019, we will see the establishment of industry-funded AI and Machine Learning MSc courses in British universities that already have strong AI research programmes. They will graduate an additional 200 AI/ML MSc students per year from 2020. By 2020 we plan to establish MSc conversion courses that will give students from non-STEM backgrounds the skills they need to work in AI.
We will also be running a campaign to increase diversity in the AI workforce. As AI-based products and services become increasingly part of our lives, it will be essential that the sector is not dominated by one gender, age group, ethnicity or culture. We may well have to think about radical new ways to increase the gender balance in AI as approaches to increase gender diversity in computing over the last 30 years have not proved effective at scale.
Innovate UK, the UK government’s innovation agency, and others will be promoting the adoption of AI through throughout British industries and businesses and, indeed, within the government itself through training and awareness raising. It has also adopted a specific mission to develop AI that will “transform the prevention, early diagnosis and treatment of chronic diseases by 2030” set by the prime minister earlier this year.
One of the biggest challenges for anyone working in AI is finding ways to share data – needed to train algorithms – in a safe and secure way. A new Centre for Data Ethics and Innovation, set up this year, with the entrepreneur Roger Taylor as the first chair, will be consulting throughout 2019 with the OAI on ways of establishing “data trusts”. The OAI will be establishing data trust case studies – yet to be picked – to develop examples of good practice that others can copy
These implementations will start to take effect in 2019, although it will be many years before we can really evaluate the impact government intervention will have on the UK’s role in the development of AI. However, one thing is clear: we will also need to collaborate internationally. AI has the potential to solve many of the biggest challenges society faces in the 21st century. But only by pooling our resources – data, research results, skills and expertise – will we be able to achieve the results we need. The UK’s new AI Council will ensure this is the case, while still enabling UK companies to compete internationally and sell the products and services its strategy will have helped them create.
Wendy Hall is regius professor of computer science at the University of Southampton
This article was originally published by WIRED UK
