Skip to content

Studies

Leila Guerra, LSE educational innovation expert: ‘Universities must teach students to judge AI’s answers’

Events Institutional 9 October 2026

‘The university must be led by people with purpose and judgement, and enhanced by machines in capacity and reach.’ That is one of the conclusions of the lecture ‘The Human-Led, Machine-Enabled University: Rethinking pedagogy, governance and societal impact in the age of AI’, delivered by Leila Guerra, Deputy Pro-Director for Education at the London School of Economics and Political Science (LSE), with which IQS has launched the IQS-LSE Talks series. The aim of this initiative, which comprises four sessions, is to explore, with expert voices from the prestigious London institution, the impact of AI in four areas: education, organisations, employment and society.

Guerra, who has more than 20 years’ experience in strategy development, change management, organisational transformation and innovation in higher education at institutions including the LSE, Imperial College London and Singapore Management University (SMU), spoke on education.

When the answer stops being what matters

Ana had three assignments due in the same week and turned to artificial intelligence to meet the deadlines. She handed them in, but when she was asked to defend her work, she could not: she remembered nothing of what she had submitted. With this example, Guerra illustrated the risk that most concerns universities today, which is not plagiarism but technology ending up thinking in place of the student. She cited a MIT Media Lab study according to which people who wrote an essay with ChatGPT showed up to 55 per cent less brain connectivity than those who wrote it unaided, as well as more difficulty recalling what they had just written.

The underlying problem, she explained, is that AI has made answers an abundant commodity. Obtaining one is no longer an achievement. What counts is what comes next: understanding why it is correct, knowing when to apply it and taking responsibility for doing so. ‘Who is accountable when we ask an agent to do a task for us?’, she wondered. Hence her distinction between cognitive substitution, in which the tool does the student’s work for them, and cognitive augmentation, in which it expands what the person is able to understand, decide and achieve.

This shift forces universities to ask what they are for. If they do not want to be reduced to degree-issuing machines, content providers or skills factories, Guerra argued, they must reclaim what no tool can offer: the chance to meet people (networking), perspective, critical thinking and personal growth through self-directed learning.

Speed with direction

Guerra reviewed how universities’ view of AI has changed. From a threat associated with plagiarism, it came to be seen as a tool for efficiency, capable of automating marking or personalising content, and today it has become a race in which each institution defines strategies and structures without anyone knowing what will come next. The risk, she warned, is mistaking speed for progress. In this regard, she stressed the importance of having a direction, one set not only by technology but also by the ability to ask the right questions and to measure the consequences. ‘STEM disciplines matter, of course, but this is the moment for the social sciences.’

Once there is a direction, how does one build a human-led, machine-enabled university? Guerra set out a four-step roadmap. The first is to decide what to delegate and what to keep. The machine can take on personalised practice, immediate feedback or realistic simulations, while people are responsible for deciding what a task is for and for answering for decisions, mentoring and ethics, which is precisely why society trusts universities. Second, she proposed rethinking pedagogy. ‘A university degree should certify the capacity for judgement in analysis. In programming, for example, the student is no longer the one who produces the code, but the one who judges it.’ She therefore argued for assessment centred on reasoning rather than on the outcome.

For the change to reach the whole institution, the speaker suggested, as a third step, governance that is strict on purpose and flexible in practice. The principles are non-negotiable: ethics, data protection, equitable access and the integrity of assessment. Within them, each discipline and each classroom should be free to decide how to use the technology, and students should take part as partners in that decision. The fourth and final step looks outwards. The university, Guerra argued, must stop being a place people pass through once in their lives and instead support them throughout their professional careers. ‘We must evolve from linear education to lifelong learning. If universities do not keep that space open, others will fill it’, she warned.

The next conversation in the IQS-LSE Talks series will take place on 18 February with Youngjin Yoo, Professor of Information Systems and Innovation at the LSE, who will address how generative AI is transforming value and organisations.