A recent report from the Policy Exchange think tank warns that the prevalence of unmonitored remote testing in universities is enabling widespread AI-assisted cheating. The findings suggest a systemic failure to update academic policies, which may be disadvantaging students who maintain academic integrity.
The 10 per cent invigilation gap in remote testing
The current state of academic assessment is characterized by a stark disparity between the adoptin of remote testing and the tools used to secure them. According to the Policy Exchange report, 78 per cent of universities contacted via freedom of information requests utilized online or remote exams in 2024. Despite this high adoption rate, only 10 per cent of those institutions employed online invigilation for every remote examination condcuted.
This gap creates a significant vulnerability in the higher education system. As the report notes, 70 per cent of these universities intend to keep using remote assessments, even as the tools available to students for dishonest assistance become more sophisticated. This suggests that the infrastructure for monitoring students away from campus has failed to keep pace with the shift toward digital testing.
Why 67 per cent of university policies ignore generative AI
The crisis is not merely technical but regulatory. The Policy Exchange report found that 67 per cent of university policies regarding remote examinations make no specific mention of generative artificial intelligence. This lack of clarity leaves a vacuum where students may not fully understand the boundaries of acceptable tool use, or where institutions lack the explicit authority to penalize AI-assisted submissions .
Professor Philip Newton, the author of the report, argues that tools like ChatGPT have fundamentally weakened the reliability of traditional assessment methods. Because these systems can produce high-standard written answers that are difficult for faculty to identify with confidence, there is a growing risk that students using AI improperly will outscore those working independently.
The 94 per cent adoption rate and the HEPI findings
The shift toward AI is already deeply embedded in the student experience. Research from the Higher Education Policy Institute (HEPI) indicates that 94 per cent of surveyed students have used generative AI in some form during their assessments. While nearly half of these students believe the technology has improved their experience, 16 per cent reported a negative impact, citing concerns that over-reliance on AI could erode their own critical thinking and learning efforts.
This trend reflects a broader tension in global education: the struggle to integrate powerful productivity tools without sacrificing the cognitive struggle essential to learning. When the majority of a student body is utilizing AI, the traditional "take-home" essay ceases to be a reliable proxy for a student's actual knowledge or skill level.
Sir Vince Cable's call for an urgent assessment overhaul
The report advocates for a return to supervised environments to restore the credibility of university degrees. Sir Vince Cable, a former cabinet minister responsible for universities, has called for an urgent and significant overhaul of how students are assessed. This sentiment is echoed by Lord Mendoza, provost of Oriel College , Oxford, who stated that unmonitored online exams and take-home essays actively encourage students to use AI, creating an unfair environment for both staff and honest students .
To combat this, Policy Exchange recommends that face-to-face assessments become the standard again. Whether these are traditional pen-and-paper tests or supervised digital exams, the goal is to ensure that the work submitted is the product of the student's own mind.
The clash between Universities UK and the Office for Students
There remains a fundamental disagreement over the role of AI in the classroom. Universities UK argues that because AI is transforming the professional workplace, institutions must teach students to use the technology responsibly rather than simply banning it. they maintain that degrees remain trustworthy because universities are developing policies to address breaches and support genuine learning.
However, the Policy Exchange report pushes for a more aggressive regulatory stance,calling on the Quality Assurance Agency and the Office for Students to classify unmonitored remote exams as "non-compliant" unless universities can prove they are secure.. This raises a critical unanswered question: how will the Office for Students define a "secure" remote exam in an era where AI can be run on secondary devices that are invisible to screen-monitoring software? Furthermore, it remains unclear how universities will balance the need for security with the accessibility benefits that remote testing provides to disabled or distant learners.
Comments 0