Universities facing a financial crisis should stop trying to ban students from using artificial intelligence and instead use the technology to test their reasoning, a higher education commentary has argued.
The warning comes as falling enrolments and reduced tuition revenue leave many institutions struggling financially, despite rising prices. Syracuse University chancellor J Michael Haynie said the situation was “a reminder that the way we have always done things is no longer adequate or enough”.
The financial pressures are already affecting students, with measures to keep universities afloat potentially leading to higher meal-plan costs, shorter library opening hours and fewer academic advisers. The commentary argues that these changes could damage an education for which students may still be paying into their sixties.
Universities face pressure from falling standards and AI use
At the same time, universities are enrolling students whose approach to learning has been shaped by generative AI. The material cited in the commentary suggests that using the technology can improve homework scores by 18 per cent while reducing monthly exam scores by 20 per cent within six months.
Regular use of large language models has also been linked to lower neural, linguistic and behavioural performance levels, it says. A study at the Massachusetts Institute of Technology described AI as “upending foundational elements of the MIT education experience”, while undermining student mastery and confidence and making it harder to assess progress.
The challenges are compounded by weaker academic performance among school pupils. The OECD’s Programme for International Student Assessment found that 15-year-olds in OECD countries had recorded their lowest average performance in science, reading and mathematics.
Against that backdrop, the commentary argues that universities are dealing with students less prepared for demanding academic work while also facing financial constraints that threaten the quality and purpose of higher education.
AI could be used to assess critical thinking
Rather than asking whether students should be allowed to use AI, universities should assume that they will find ways to do so, the commentary says. The focus should instead return to developing cognitive ability and preparing students for life and work beyond university.
That would mean placing less emphasis on written outputs, which AI can produce quickly, and more on whether students understand their work, can explain their logic and are able to defend an argument.
Employers cited in the commentary placed greater value on critical thinking and problem-solving, at 73 per cent, and adaptability or learning agility, at 65 per cent, than on AI and related technical skills, which were sought by just under half of hiring managers.
Traditional teaching methods already offer a model for testing those abilities. Humanities students at Oxford and Cambridge write weekly essays and discuss them with tutors, while students on France’s classes préparatoires take part in regular oral examinations designed to test rapid problem-solving and depth of knowledge.
Such approaches are intensive and costly, making them difficult to provide at scale. The proposed alternative is for AI to hold short weekly conversations with students about how they reached their conclusions, identifying those who need further help from a lecturer.
The aim would be to use AI to support, rather than replace, teaching. Examples cited include an AI agent at the University of Wollongong providing first-year guidance from educator-approved material, and an AI-powered robot helping trainee teachers practise handling difficult conversations with parents.
The commentary argues that universities must adapt both to changing student behaviour and to the financial pressures affecting the sector. Combining technology with more rigorous assessments of reasoning, it says, could help institutions justify their costs while producing graduates better prepared to use AI responsibly in the workplace.
