Higher Education: Affordability and Quality

Lord Tarassenko Excerpts

Thursday 2nd July 2026 Lords Chamber

[Lord Tarassenko (CB)]

My Lords, I congratulate my noble friend Lady Deech on securing today’s debate. Her commitment to the issues of quality and affordability in higher education is second to none. I declare an interest as the founding president of Reuben College, Oxford’s newest college: a 21st century college in a 900 year-old university.

In early 2025, just over two years after the launch of ChatGPT, a survey carried out by the Higher Education Policy Institute revealed that 90% of UK undergraduates were using AI in their assessments. As a professor from the University of Warwick recently put it,

“all typed work by students is now worthless as a means of assessing them and their knowledge”.

Many departments are going back to hand-written assessments for all tests that matter, and I believe that we should contemplate reintroducing all exams as part of summative assessments, such as finals. Right now, conversations about AI in the higher education sector are mostly defensive, and this needs to change. The question should not be “How do we catch students cheating with AI?” but “How can we reimagine the delivery of education in the age of AI?”

As with the social media ban, it appears that Australia is ahead of the UK in this domain. In a recent paper, two Australian academics argued that, as AI automates analytical tasks, the true purpose of a university must shift from teaching students how to generate answers to developing capabilities in reasoning, critical judgment and acting under uncertainty.

My estimate is that we need only about 20% of the lectures which we required prior to the advent of generative AI. Lectures should be given only to introduce a subject, occasionally go through a particularly difficult topic or proof, and then provide a synthesis at the end. In between, students can—and will, in any case—use their favourite generative AI tools to progress, at their own pace and with immediate, personalised feedback, through the subject framed by the lectures. The time which professors would save by giving only 20% of the pre-AI age lectures should be devoted to their facilitating small-group seminars, typically with 10 to 12 undergraduates in each group. In these settings, the students would not be allowed to have their smartphones or laptops with them. The professor would probe their knowledge and enable them to develop the critical cognitive capabilities now being blunted by the regular use of AI during the learning process.

Of course, small-group teaching, as opposed to giving lectures to 200 or 300 students, is much more inefficient. For such a strategy to be feasible, the higher education sector will need to embrace the use of graduate teaching assistants. As in the US, these would be PhD students or post-doctoral researchers, who would be trained to deliver high-quality, small-group teaching, working in tandem with the professor responsible for the undergraduate course.

The Department for Education cannot tell universities how to teach, but it controls the financial levers. The large-scale deployment of graduate teaching assistants to support small-group teaching would require DfE and UKRI to develop plans for an alternative model for funding PhDs. Those UK-domiciled postgraduate students who chose the option to become teaching assistants would be able to start paying off their student loan while doing their PhDs.

I am aware that the Office for Students has launched a research project to investigate how universities are using AI, with a nationwide data-collection phase closing at the end of next week. The OfS says that it will

“analyse the findings in collaboration with Advance HE”

and that it expects

“to publish the findings later this year to support institutions in using AI tools to improve students’ experiences of higher education”.

But we need to be bolder in our ambition if we are to maintain the quality of higher education in the age of AI. We need to start now to develop the rigorous training programmes required to turn PhD students and postdocs into expert facilitators, able to lead device-free sessions in which undergraduates can develop genuine critical thinking and independent reasoning.

AI-guided learning, framed by just a few lectures and combined with rigorous oral probing in small-group seminars, is surely of higher quality than 100 hours of sitting passively in a lecture theatre. We need the DfE to fund pilot programmes in selected university departments, from the humanities to engineering, to demonstrate how undergrad teaching can best be adapted for the age of AI. It is a challenge which we cannot ignore, and I hope that the Minister will be able to reassure the House that the DfE is devoting some critical thinking time to addressing the issue.