The debate over artificial intelligence in university admissions has crossed the Atlantic. At Stanford University, students are locked in a fierce argument over whether algorithms should help select the next generation of undergraduates. Now British universities are watching closely, weighing their own ethical frameworks as they quietly explore similar tools.
Stanford’s undergraduate senate last week held a heated session on the issue. Some students argued AI could reduce human bias in sifting through tens of thousands of applications. Others warned it would embed existing inequalities deeper into the system, punishing applicants from state schools or non-traditional backgrounds. The university has not adopted AI for admissions, but the discussion has become a flashpoint for wider anxieties about technology’s role in education.
Across the pond, the debate lands on different ground. British universities already use contextual data to adjust offers for disadvantaged students. But AI promises to go further: scanning personal statements, predicting dropout risk, even automating the first read of an application. The Russell Group confirmed to this paper that several of its member institutions are “exploring the ethical use of machine learning in admissions” but stressed no decisions have been made.
Dr. Harriet Walsh, a lecturer in digital ethics at the University of Manchester, told me the stakes are high. “We have to ask: what are we optimising for? If an algorithm learns from past successful applicants, it will replicate past biases. That means fewer working-class students, fewer from ethnic minorities, fewer from the North. AI doesn’t fix the problem. It just automates it.”
Student groups are mobilising. The National Union of Students this week launched a campaign called “Fair Access, Not Fast Access” demanding a moratorium on AI in selection. “This is about who gets a shot at a degree,” says NUS vice-president Amara Okafor. “We have seen what happens when tech companies make these calls. They make mistakes. And the mistakes fall on the poorest.”
But proponents argue AI could be a force for good. Professor James Hartley of the University of Bristol, an expert in computational social science, believes well-designed algorithms can spot potential that human admissions officers miss. “We are drowning in applications. AI can flag candidates from schools where few go to university, measure resilience from their personal statement language, or even cut down on the sheer time it takes to read every file. That frees up staff to do the human judging for borderline cases.”
The Department for Education declined to comment on specific university plans but said it is “monitoring developments closely” and expects any use of AI to be “transparent and fair”.
Yet the real test may come from students themselves. At Stanford this week, a petition demanding a full ban on AI in admissions gathered 2,000 signatures in 48 hours. Similar petitions could emerge here. The British admissions system is already under strain from grade inflation, rising applications, and the cost of living crisis that makes every offer a lifeline. Adding AI into the mix without public consent feels reckless.
What happens at Stanford matters here. If that flagship debate ends in a ban, it will stiffen spines among British vice-chancellors. If it proceeds, the pressure to keep up will be immense. Either way, the question is no longer whether AI will enter admissions. It is whether we will control it before it controls the door to our universities.
The price of getting this wrong is measured in futures. And futures, unlike algorithms, do not come with a reset button.
