A University of Reading study found that AI-generated exam answers often go unnoticed by experienced exam markers, with 94% of these answers going unnoticed and receiving higher grades than student submissions. Researchers are calling on the global education sector to develop new policies and directions to address this problem. The study highlights the need for sector agreement on the use of AI in education and highlights the responsibility of educators to maintain academic integrity. The University of Reading is already taking steps to integrate AI into teaching and assessment to better prepare students for the future.
Research from the University of Reading shows that AI-generated responses often escape detection in academic assessments and can outperform student responses, calling for a global update of educational policies and practices in AI material.
Researchers have found that even experienced exam markers can struggle to identify answers produced by artificial intelligence (AI). The study, conducted at the University of Reading in the United Kingdom, is part of an initiative by university administrators to assess the risks and benefits of AI in research, teaching, learning and assessment. As a result of their findings, updated guidelines have been distributed to faculty and students.
The researchers call on the global education sector to follow the lead of Reading and others who are also developing new policies and guidance and do more to respond to this emerging problem.
In a rigorous blind test of a real university examination system, recently published in PLOS ONEChatGPT generated exam answers submitted for several undergraduate psychology modules that went undetected 94% of the time and averaged higher marks than actual student submissions.
This is the largest and most robust blinded study of its kind, to date, aimed at challenging human educators to detect AI-generated content.
Study results and educational impact
Associate Professor Peter Scarfe and Professor Etienne Roesch, who led the study at Reading’s School of Psychology and Clinical Language Sciences, said their findings should be a “wake-up call” to educators around the world. A recent UNESCO survey of 450 schools and universities found that less than 10% had policies or guidance on the use of generative AI.
Dr Scarfe said: “Many institutions have moved away from traditional exams to make assessment more inclusive. Our research shows that understanding how AI will affect the integrity of educational assessments is of international importance.
“We won’t necessarily return entirely to handwritten exams, but the global education sector will need to evolve in the face of AI.
“It is testament to the frank academic rigour and commitment to research integrity at Reading that we have turned the microscope on ourselves to lead this process.”
Ethical Considerations and Use of AI
Professor Roesch said: “As a sector, we need to agree on how we expect students to use and recognise the role of AI in their work. The same goes for the wider use of AI in other areas of life to avoid a crisis of trust in society.
“Our study highlights the responsibility we have as producers and consumers of information. We must redouble our commitment to academic integrity and research.
Professor Elizabeth McCrum, Deputy Vice-Chancellor for Education and Student Experience at the University of Reading, said: “It is clear that AI will have a transformative impact on many aspects of our lives, including how we teach students and assess their learning.
“At Reading, we have undertaken an extensive program of work to consider all aspects of our teaching, including making greater use of technology to improve the student experience and strengthen graduates’ employability skills.
“Solutions include moving away from outdated ideas about assessment and adopting models that are more tailored to the skills students will need in the workplace, including through the use of AI. It is essential to share alternative approaches that allow students to demonstrate their knowledge and skills with colleagues across disciplines.
“I am confident that with the detailed analysis that Reading has already carried out across all of our courses, we are able to help our current and future students discover and benefit from the rapid developments in AI.”
