Written coursework for A-level and GCSE exams in the UK will undergo significantly more scrutiny to combat the use of artificial intelligence, according to Ofqual, the nation's exams regulator. Sir Ian Bauckham, Chief Regulator at Ofqual, stated that reforms to these qualifications must include safeguards against "AI fraud," particularly for "extended writing coursework," which is considered especially vulnerable to cheating.
These measures are being developed as part of the Government's redesign of qualifications for 16-18 year olds, following the Curriculum and Assessment Review. New content, subjects, and a reduced emphasis on timed exams at GCSE level are slated to be introduced from 2029. Additionally, new vocational qualifications, V levels, will be rolled out starting in 2027.
In a speech at the Festival of Education, Sir Ian emphasized that AI-generated output should not be considered a substitute for "genuine human endeavor." He expressed his disagreement with the notion that allowing students to use AI for assessments is beneficial, noting that many universities, including UCL Law School, are increasing the use of in-person, invigilated exams to "AI-proof" their assessments.
"Offloading student thinking to AI does not develop the human mind," Sir Ian is expected to say. "It leads ultimately to less learning, less satisfaction, less usefulness in the workplace, and ultimately is a covert assault on the realisation of human potential."
Sir Ian believes that protecting the integrity of England's qualifications system is vital to maintaining its global respect as a "national asset." He rejected the idea that future generations will require fewer qualifications because AI will perform most tasks, arguing instead that the future workforce will need higher-level expertise to augment AI capabilities. "We won't need less education, we will need more of it, and we will need qualifications to frame it," he stated.
These concerns are amplified by recent findings from a Higher Education Policy Institute poll. The survey of 1,000 university students revealed that 95 percent admitted using AI tools like ChatGPT for assessments. Furthermore, 12 percent reported using AI to generate text for submission, an increase from previous years. While nearly half of students felt AI improved their experience, 16 percent found it detrimental, with some complaining that it contributes to laziness and negatively impacts learning and skill development.