The widespread adoption of artificial intelligence (AI) tools is significantly transforming the university experience, raising concerns about students’ preparedness for the job market and the integrity of academic assessments. A recent survey by the Higher Education Policy Institute (HEPI) found that 95% of university students have used AI in their studies, reflecting a deep integration of these technologies in academic work.
This trend comes at a time when youth unemployment in the UK has reached its highest level since 2013. Some students and academics worry that heavy reliance on AI may undermine critical thinking skills, subject knowledge, and the ability to perform in professional settings. One student preparing to start a master’s degree in aerospace engineering admitted that their use of AI during undergraduate studies had weakened their grasp of fundamental concepts, particularly in mathematics. The student, who is expected to graduate with first-class honors, expressed concerns about facing induction tests for future employment, fearing that an overdependence on AI might hinder their performance.
Alongside student worries, a 2025 survey by educational charity Jisc highlighted student anxiety regarding how AI might diminish their capacity for developing ideas, thinking critically, and communicating effectively. Some educators are exploring ways to incorporate AI into teaching, with David Andersen, an associate professor at Durham University, noting that some colleagues encourage students to engage with AI in debate formats as a pedagogical tool. “It is becoming an extension of the teacher and a tool for the students,” he said.
Despite attempts by some students to use AI responsibly—such as for research assistance, verifying answers, or expanding upon their own ideas—there is widespread misuse as well. The same aerospace engineering student admitted using AI for exam questions under time restrictions and employing the technology to compare their dissertation with an AI-produced version, noting that while it would have been tempting to submit the AI draft, they chose not to.
Challenges in detecting AI-assisted academic dishonesty are growing. Ofqual’s chief regulator recently acknowledged difficulties in identifying AI use in A-level exams, particularly in subjects reliant on extensive written coursework. The current exam formats, often based on recycled practice papers, appear ill-equipped to counter AI’s capabilities. The student noted that many peers reportedly answer a significant majority of exam questions easily using AI prompts.
A report published last week by Policy Exchange underscored the need for universities to adapt assessment methods to the realities of AI. It found that 78% of UK universities currently utilize online or remote examinations, with 70% planning to continue this practice, yet only 10% employ online invigilation techniques. Andersen criticized the slow pace of innovation in academic assessments post-pandemic, emphasizing that merely moving exams online has not effectively addressed the challenges posed by AI. “We need to innovate and change in a system that has resisted rapid change for a millennium. We are not moving fast enough and, honestly, I struggle to envision how we can,” he said.
As AI tools become increasingly sophisticated and prevalent, the higher education sector faces mounting pressure to rethink teaching and evaluation methods to ensure students develop genuine skills and knowledge necessary for future employment.