There is no question that artificial intelligence (AI) and automation are entering the workplace in many graduate level jobs, and this trend is likely to continue and quicken.

Times Higher Education recently asked whether universities needed to rethink what they do and how they do it, given that artificial intelligence is beginning to take over many post-university careers.

The implications and – most importantly – the potential benefits for education are significant, and perhaps not yet appreciated by higher education leaders.

With that in mind, here are four examples of how AI can benefit universities.

HE is ideally placed to prepare students for the AI world

First, there is a new role for HE to equip graduates to work effectively alongside artificially intelligent systems.

The onslaught of AI on white-collar jobs is likely to lead to the AI augmentation of human intelligence, rather than the total replacement of human workers with machine workers. We need workers who understand how to make the best use of the power that AI automation can bring to industry and commerce.

Workers who know where and how human intelligence can work with AI to achieve this augmented human intelligence – an intelligence that is greater than the sum of its parts – can increase productivity. HE could and should lead the way in developing its educators’ knowledge and skills about AI and the ways that people can work effectively with AI systems.

University lecturers, therefore, must then ensure that students enter the workplace with an understanding of AI’s competencies and limitations, and the capability to continue learning as AI develops and continues to change workplace roles and expectations.

To ignore this opportunity brings with it a real danger that the economic and social benefits of AI will be squandered, because we have invested billions in the cost of automation, but nothing to ensure that the workforce – most notably the educators – are equipped and skilled enough to take advantage of what the fourth industrial revolution can provide.

AI can help solve the big HE challenges

Second, AI can provide some of the solution to the big challenges faced by higher education. For example, there has been an enormous rise in the amount of educational data about students that is available to universities.

Numerous articles (like this, and this) have highlighted the data patterns about student activity that are routinely recorded and analysed by universities using “learning analytics”. At the moment, these analytics are mainly used to predict which students might be about to fail or drop out, but the potential for AI informed Learning Analytics to analyse the learning processes of an individual student and to provide timely interventions in real time to support that student are enormous.

To date, AI technologies for education are underutilised and underfunded and yet they hold a big part of the answer to how we harness AI to improve HE.

Researchers can be a key part of the AI revolution

Thirdly, HE researchers and educators are the key people to help the UK to develop the right type of AI systems for use in education, and they need to be doing this right now.

We need AI systems that move beyond the machine learning and neural network techniques that dominate the work of the main AI protagonists within and beyond education, such as “robot tutor” Knewton and Google’s game-playing algorithm, DeepMind.

To tackle the really big challenges within education, we need AI systems that can explain their reasoning, justify their decisions and negotiate with their users. We would not employ a human educator who could not explain what they were doing and why. So why would we be willing to accept anything less from machines?  Investment in HE collaborations with the edtech sector is the essential ingredient in helping UK HE to drive the development of these systems.

Universities are well placed to research AI

Fourthly, there is a role for HE to conduct more fundamental research into the AI systems that will improve the future of education for all students.

We know that there are essential human abilities and skills that AI is a long way from mastering: social intelligence, empathy, love and creativity, for example. In addition to these essential human abilities there is another that is crucially important to education: the human ability to contextualise our actions.

Context is a key variable in the effectiveness of education; some would argue it is the biggest variable in a student’s propensity for success.

Context can be defined as the combination of people, environments, knowledge, technology and resources that are part of every student’s interactions with the world and that are the significant variables that impact upon each student’s educational success. AI developers have yet to master the kind of contextual modelling that would enable AI to appreciate how and why what works in a university in Salford, may not work at a university in Surrey, Seattle, Sydney or Singapore.

Whilst it may be possible to use machine learning and neural network technology to build systems that can adapt to contextual factors, these systems will need modelling solutions to enable them to justify the way they have used these contextual factors to make decisions. These modelling techniques do not yet exist and along with much of what AI will offer, they raise significant and important ethical questions that must be addressed as a matter of urgency to avoid hindering the progress of AI benefits for HE.

Higher education must rise to the challenge of both the contextual modelling research and the essential ethical debate.< Put simply, AI has the potential to bring great benefits to HE; most notably to graduates and postgraduates. However, we will only reap these benefits if we develop and use AI as the innovation that it is and with the vision and energy that students deserve. >Author Bio: Rose Luckin is chair of learning with digital technologies at the UCL Institute of Education.