University Lecturer's AI Writing Experiment Backfires as Students Prefer ChatGPT
Students Prefer ChatGPT Over Lecturer's Writing in University Test

University Lecturer's AI Writing Experiment Yields Unexpected Results

An academic's attempt to demonstrate the limitations of artificial intelligence in creative writing has produced surprising and uncomfortable findings. Sarfraz Manzoor, a non-fiction writing lecturer at University College London and established writer of three decades, designed a classroom experiment that challenged his assumptions about AI's capabilities.

The Anxiety of Modern Authorship

Professional writing has always been accompanied by anxiety, according to Manzoor. Throughout his career, concerns have evolved from securing commissions and financial stability to more contemporary fears about technological displacement. "These days, I still have all these worries, but they have been joined by a more recent anxiety: that I will be replaced by AI," he acknowledges.

Despite widespread warnings about artificial intelligence's threat to creative professions, Manzoor maintained confidence that AI could not replicate the distinctive qualities of human writing. To prove this to his students, he devised a comparative test between his own work and content generated by ChatGPT.

The Classroom Comparison

Manzoor presented his UCL students with two pieces formatted identically. The first was his own column reflecting on finding comfort in predictable elements during the unpredictable year of 2026. The second was generated by ChatGPT, specifically prompted to produce 700 words "in the style of Sarfraz Manzoor" making the same argument.

The lecturer anticipated his students would easily recognise the superiority of human writing, appreciating its specificity, wit, and rhythm. Instead, the majority expressed preference for the AI-generated version, describing it as better argued, more clearly structured, and more ambitious in scope. Most painfully for Manzoor, some students even found the AI piece "more personal" than his own.

Cultural Disconnect and Educational Priorities

When Manzoor revealed which piece was human-authored, students appeared genuinely shocked. Their subsequent explanations revealed important insights about contemporary education and cultural literacy. Many were international students, predominantly Chinese, who found the cultural references in Manzoor's piece inaccessible.

References to British cultural touchstones like Pulp's Disco 2000, Bonnie Blue, and Nigel Farage created barriers rather than connections for these readers. Furthermore, students had been educated to value writing with clear structural signposting and smooth transitions between points - qualities they found more prominently in the AI version.

Diverging Perspectives on Quality

The experiment revealed a significant divide in how different readers evaluate writing quality. Students with greater media literacy - particularly those who subscribed to newspapers - expressed stronger preference for Manzoor's original work, describing the AI version as "cold", forgettable, and temporally ambiguous.

One student quoted a memorable line from Manzoor's piece, while another was inspired to listen to Disco 2000 for the first time. These responses suggested that while AI writing offered competent structure, human writing provided richer rewards for engaged readers.

The Future of Writing and Education

Manzoor reflects that AI-generated content creates an illusion of profundity through bland banalities that might superficially resemble insight. While this satisfied some students, he doubts it would convince professional editors, who would likely request more authentic authorial voice and personal connection.

"My homework for the rest of term is to teach my students not only how to be better writers, but how to recognise and value writing that is truly human," Manzoor concludes. The experiment has reinforced his belief that writing's essential qualities - distinctive voice and cultural specificity - remain valuable precisely because they're not universally immediately appreciated.

The classroom test serves as a microcosm of broader debates about AI's role in creative fields, suggesting that while artificial intelligence can produce competent imitations, human writing's unique value persists in its capacity for genuine connection and cultural resonance.