Solving the UK’s AI skills gap

Personal perspective

Accenture logo

Authors

Ray Eitel-Porter

Ray Eitel-Porter

Global Responsible AI Lead at Accenture


Addressing the UK’s AI skills gap has never been more important. COVID-19 disrupted business-as-usual for most companies, and many turned to digital technology, including AI, to maintain operations. Research by Accenture for its Technology Vision 2021 report found that 34% of UK organisations have scaled up AI technologies in response to COVID-19. However, it also found that only 27% of UK business leaders think their non-technical workforce is well-prepared to leverage new technology.

Just as the UK needs people with AI skills more than ever, the skills gap is broadening, particularly for groups that were already underrepresented. For instance, as illustrated in The Alan Turing Institute’s report, Where are the women? Mapping the gender job gap in AI, extensive disparities between women and men persist in skills, status, pay, seniority, industry, job, attrition and educational background.

It is against this backdrop that the Turing convened a webinar on how the UK can address issues around a lack of skills and diversity in AI, beyond the work already completed by the AI Council and reflected in its excellent AI Roadmap

For the debate, I was joined by Professor Charlotte Deane, Deputy Executive Chair, EPSRC and COVID-Response Director UKRI; Maggie Philbin, Co-Founder of TeenTech; Priya Lakhani, CEO of CENTURY and UK AI Council Member; and Matthew Forshaw, National Skills Lead at The Alan Turing Institute. The session was moderated by Dame Wendy Hall, Scientific Advisory Board Member at The Alan Turing Institute. This article outlines some of the main themes discussed.

The scale of the challenge

The skills gap facing the UK runs across the entire learning journey, from early years through to senior positions in academia and industry. For example, only a fraction of the number of doctrinal students required by research and academia is currently being secured. The lack of diversity is similarly concerning. Currently, just 22% of people in AI roles are female. 

While some figures are available, there is in fact a lack of rigorous data around the regional and national skills gaps being experienced in the UK. Without this data, it is difficult to plan effectively, and some organisations may fail to grasp the true scale of the challenge facing the country. Getting this data and providing it to policymakers and training providers is a priority. 

Addressing diversity

Solving the lack of diversity in AI will help address the overall skills gap by broadening the pool of available talent. It will also lead to better work, as people with different experiences and outlooks are needed to create diverse solutions.

Diversity needs to run through all levels of seniority. At present, embedding diversity in senior levels is a challenge, and one that needs to be addressed through targeting lifelong learners and ensuring that people from diverse backgrounds joining at a junior level are retained. 

The culture and ethics of AI

Data skills delivery is about more than technical skills alone. What’s required is wholesale cultural change. Within businesses and other organisations, the optimal use of AI will see a shift from traditional operations-led services to data-led services, where data and analytics drive business strategy. It also requires trust on behalf of the owners of the data.

Ethics and public trust must therefore be integral to data skills initiatives. In this regard, it’s good to see that the ethical, safe and trustworthy development of responsible AI will form a big part of the Government’s National AI Strategy, due to be published later this year. 

The professionalisation of AI

There is currently no consistency in the terminology used to describe the knowledge, skills and roles of data science and AI. This makes it hard for employers to assess the skills content of university courses or to be consistent in their job advertisements. Discrepancies in definitions can also lead to the inconsistent adoption of best practice and uncertainty among end users. 

Strong leadership is required in shaping curriculum standards and accreditation, drawing from the expertise of the broad range of bodies within the field. In this respect, there are lessons to be learned from how the government is approaching adjacent fields. For example, the UK government is creating a new organisation, the UK Cyber Security Council, which will act as the cyber security industry’s professional standards and development oversight body. A similar organisation for AI could also make a valuable impact. 

The Turing, as part of the Alliance for Data Science Professionals, is working alongside professional bodies (including the Royal Statistical Society, BCS, Royal Academy of Engineering) to explore the role of accreditation to ensure that educational programmes deliver the skills needed by people and industry. The aim is also to raise awareness of, and access to, routes from academic and vocational training into professions. 

AI learning in schools

The AI Council has recommended an online academy for understanding AI, which would provide trusted materials and initiatives to support teachers, school students and lifelong learning. This work will help bridge the gulf between what is currently taught in schools and the future skills that people will actually need. Ideally, it will also help people take direct control of their own learning by providing a clear source of information and guidance. 

But more will need to be done, including significant changes to the curriculum. Teachers are already stretched, and the national curriculum full to capacity. Therefore, rather than adding an additional course, data literacy skills should be embedded across a range of existing subjects. It would be useful to have an AI curriculum guide, developed by experts, offering lesson plans and advice to teachers. In addition, the Free Schools model could be used to create flagship AI schools led by academia and industry. 

Another promising approach would be to identify exceptional teachers and academics who face barriers when it comes to AI teaching in terms of accessing resources or understanding the pedagogy. By training the trainers through mentoring and best practice sharing they can be given a boost that will be passed on to their students. 

Finally, it is important that further education isn’t forgotten. There is a significant opportunity to create AI-literate workers by offering bespoke digital programmes at colleges that can be selected as an elective module by students. 

Boosting skills in the short-term

Changes to schooling and further education will have an impact in the medium- to long-term. But to drive short-term change, we need different solutions. One approach could be to incentivise large businesses to run training schemes and apprenticeships. Through churn, a number of these apprentices would also go on to work for startups and smaller companies.

Another approach would be to proactively support women who are returning to work following maternity leave. Removing any possible bias against them would immediately boost the talent pool. 

Employers also have a role to play, as they can help people already in the workforce reskill for a career in AI. Accenture provides one such programme, offered to people who studied relevant data-related topics in the past and now want to re-enter the field. Such efforts would be greatly enhanced with the addition of a national online skills academy. 

A concerted effort

A great deal remains to be done to solve the UK skills gap. However, the webinar panel agreed that a solution is achievable. What’s required now is for academia, industry and government to work together to put existing ideas into practice and to think of new ways to solve the challenge. Finding an answer will provide a significant boost for the UK while providing a rewarding career for many.