Universities and tech companies can collaborate to create a better workforce
The director of social mobility at King’s College London says that institutions and industry can work together to narrow attainment gaps
King’s College London’s motto is “in service of society”. That is the lens through which it examines everything, including the role of artificial intelligence in supporting social mobility. Anne-Marie Canning, director of social mobility and student success at King’s, advocates human judgement and contextualised processes used alongside the power of machine learning in order to reach under-represented groups.
“As King’s College London matures to be 200 years old, we have a clear strategy outlined in our 2029 vision. One of the plinths is we want to be the top Russell Group university for social mobility,” she says. This ambition is framed in terms of “offering a transformative education with academic outcomes and fullest participation in the labour market”.
The past five years has been about King’s building diversity into its student recruitment. Its future focus is maintaining student success and closing attainment gaps. “Gaps exist in all shapes, sizes and forms, in education and labour markets,” explains Ms Canning. “Universities can’t solve these on their own, [they] need major innovators and employers to think about how they recruit and to understand talent and its potential.
“You have to know how to look beyond the usual suspects in terms of recruitment practices – particularly in technology, where we know that diversity is critical for creating innovation. Creating diverse workforces has to be about partnership: universities have been good at partnering with schools and colleges, but it’s about how we team up with employers going forward.”
A partnership that combines the innovation of tech companies with the best practices of universities may narrow attainment gaps. “Our work with The Bridge Group around social mobility with large employers reveals an acknowledgment in the corporate sector that universities are more advanced in understanding what talent looks like and better at identifying it,” she says.
Universities are 10-15 years ahead of the business world in their use of contextualised recruitment processes, says Ms Canning. “At King’s, admission is not just about giving out prizes for good grades…recruitment is based on a student’s potential to benefit from a course and to positively change the world. We are recasting the admissions process.”
Digital technology and AI are introducing new opportunities to further increase access. Ms Canning cites the recruitment system she saw at a major beauty corporation as an example. “The system was situational and for some students this might be an environment in which they perform better than a sit-down interview or pre-screen. Using virtual reality, the candidate was put into a boardroom space and asked to respond to situations.”
Algorithms may be getting a bad name in some sectors for perpetuating bias on an industrial scale, but teamed with safety checks and human judgement, AI has the potential to be a force for the good, says Ms Canning. “The danger [of systemic bias] is not a new problem. Exams’ marking schemes and rubrics bring the same pitfalls in entrance examinations. It’s all about how you design in diversity with the algorithm.”
King’s uses algorithms and machine learning to recruit for some courses. This digital processing is accompanied by manual safety checks. “What we learnt from the admissions process is that you constantly have to review data and see who is winning and who is losing. Systems can easily be thrown to advantage or disadvantage certain groups.” With diversity, there’s always an outlier that a systemised approach would not recognise, she adds.
“The rubric says ‘no way’ to a predicted B-grade candidate, whereas a briefed human reviewer looks at an application and sees they have not had a maths teacher for six months. At that point no algorithm is going to understand that contextual piece of information. It’s about bringing human judgement alongside some clear rubric. You can use algorithms to speed things up, but ultimately it takes time to get diversity right.”
Anne-Marie Canning is director of social mobility and student success at King’s College London and was part of the panel discussion “How is the fourth industrial revolution redefining human talent?” at the 2018 Huawei Academic Salon.
Learn more about Huawei in education.