AI in higher education

Artificial Intelligence “AI” has become one of the key topics of 2023, and higher education is not exempt from this wider public discourse and development.  This blog highlights some themes from the Talking HEds conversation between Gary Attle and Siân Jackson that you can watch here.

Some of the main concerns in the sector about AI include whether it is a threat to academic integrity; how to harness the technology in a way that enhances independent student learning and development; how to safeguard against ‘essay mills’ and the use of the technology to cheat. It remains to be seen what level of input the OfS might have, but there has already been some helpful guidance from the Russell Group, and we expect that there will be some useful information for the sector following on from the Government’s call for evidence on Generative AI in Education (which closed on 23 August). HE institutions that have not already done so should have a fresh look at their academic regulations to see what, in terms of the use of AI, would be permissible and what would not.

The use of AI is already notable in an employment context, and it is likely to increase further in areas like task allocation and performance management (insofar as the employee has a role that can be measured in such a way); surveillance and monitoring; and in recruitment- namely filtering CVs and automatic scoring. It is important to note that whilst there are currently no UK laws that are specifically in place to deal with AI in the context of the workplace, existing legal frameworks such as employment law, equality law and data protection law still apply.

In terms of areas to note regarding AI in the HE workplace:

• Academics and their students will need the tools to understand the proper uses and limitations of AI systems in learning, teaching, assessment and research.  This approach will need to be underpinned by a combination of academic, legal and technical considerations.
• As AI inevitably plays an increasing role in teaching, there is a question mark in terms of how this might impact on the performance appraisals and management of academics.
• By extension, there is a wider need to consider when it is appropriate for AI to be used or not used in education. This includes the possible uses in terms of student assessment and how to ensure that rigorous scholarship standards are maintained. 
• AI has the potential to speed up research processes and improve productivity. However, this needs to go hand in hand with reviewing and updating any research codes of conduct, to take account of new areas of potential research misconduct that might arise as a consequence of AI input.
• The intelligence and capability of AI programs and language models is constantly developing, but as with any program designed by humans, there are limitations and flaws. One of the flaws that can arise in practice is that the way in which a program or model is designed or trained, could be inadvertently exhibiting bias.  AI systems can also generate “hallucinations”.  Amongst other matters, equality and data protection laws will need to be taken into account, and thought given to questions of ownership and confidentiality of both the inputs and outputs to AI systems.
• An employment contract is technically a personal service arrangement (that inherently involves mutual obligations between humans and, normally, employment organisations that will be staffed by humans), so there are obvious limitations for how much an AI tool can offer in terms of decision making and accountability. Given the technical ability of AI to make or at least inform decision making, there may therefore be some potential conflict with these norms of an employment relationship.

The prevailing theme here is that HE institutions can use AI to augment their capabilities; enable them to focus on higher level tasks; and streamline teaching and research. Naturally, AI needs to be used sensibly (noting some of the considerations highlighted above), and in adherence with any applicable internal policies and any OfS or Governmental guidance.