Universities and industry need ‘common language’ around AI

Universities have been urged to work with industry to create a “common language” around artificial intelligence and establish a more consistent approach to training graduates for developing industries.

Speaking at the Times Higher Education Digital Universities UK event, Matthew Forshaw, senior adviser for skills at the Alan Turing Institute, argued that AI skills among graduates are currently “fragile and uneven” and there was a need to professionalise degrees that lead to careers in this area.

Forshaw noted that the Horizon scandal, which saw more than 900 subpostmasters wrongfully convicted of theft due to faulty IT software, highlighted the need for “strong, professional competence, and ethical activity around the development and deployment of IT systems”, which should also be extended to data scientists and AI professionals. 

Other sectors have a “far more ingrained culture of professionalisation, accreditation of degrees, and chartership of professionals than some of the feeder disciplines for data science and AI”, he noted. 

Forshaw, also a reader in data science at Newcastle University, described this as a “coordination problem” between universities, industry and government.

Employers often “define and describe AI skills very differently”, and while universities are rolling out “really exciting new programmes”, he said they are doing this “without common shared reference points for what are the skills that industry needs”.

Forshaw told delegates at the event in Birmingham: “What this really calls for is some common language and some shared terminology that universities can use when they’re developing their programmes so they can articulate what’s on offer; that students can use to self-assess where they are in terms of their capabilities versus the expectations based on these values; and that industry organisations can use to support their hiring practices and support recognition. 

“At the moment, we have this gulf in terms of differing definitions, and we see this widening and widening when we look at particular sectors, especially those that are at lower levels of maturity.”

Universities still have a “really incredible and very valuable role to play” in such a system by “convening across industry, and government in terms of being able to operate with public trust and in terms of being able to take this in a long-term way”.

“In practice”, this would mean “co-designing this capability” rather than just “responding to demand”, embedding competency frameworks into degrees and professional development courses, to “support lifelong learning at scale”.

“Universities have this opportunity to not just think about talent, not just think about our conventional pathways of crafting, training and graduating students, but more thinking about our role as a university being able to support the development of this infrastructure.”

This is “particularly” valuable in the AI space, said Forshaw, as “the learning never ends”. 

“Increasingly, these skills are developed outside traditional university environments, so I would like to see a shift in the question from, ‘how do we produce more AI graduates?’ towards ‘how do we enable AI capability across the entire workforce?’”

juliette.rowsell@timeshighereducation.com