The Human Touch: Why AI Still Can’t Match Student Writers
A groundbreaking study reveals that whilst artificial intelligence excels at grammar and structure, it fundamentally lacks the personal engagement that makes human writing compelling
There is barely a sector in today’s developed societies in which artificial intelligence has not made an impact and the realm of education is no different. It is in this space that a critical question has emerged that keeps educators awake at night: can students simply use ChatGPT to write their essays, and would anyone be able to tell the difference? A comprehensive new study from the University of East Anglia suggests the answer is more nuanced—and reassuring for educators—than many might expect.
The Great Writing Showdown
The research, published in the prestigious journal Written Communication, represents one of the most systematic comparisons to date between human and AI-generated academic writing. Professor Ken Hyland and his team at UEA’s School of Education and Lifelong Learning analysed 290 essays in total—145 written by real university students and an equal number generated by ChatGPT-3.5.
The findings were revealing: whilst AI-produced essays were generally well-structured and grammatically correct, they lacked a key element that distinguished them from authentic student submissions—the human perspective.
“Since its public release, ChatGPT has created considerable anxiety among teachers worried that students will use it to write their assignments,” explains Professor Hyland. The fear, he notes, is that such tools might facilitate academic dishonesty whilst simultaneously weakening core literacy and critical thinking skills—particularly problematic given the current absence of reliable AI detection tools.
The Missing Element: Engagement Markers
The study focused specifically on what researchers term “engagement markers”—the rhetorical devices that writers use to connect with their readers. These include rhetorical questions, personal asides, direct appeals to the audience, and commentary that reflects the writer’s stance on their subject matter.
The results were striking: human writers consistently employed a rich array of engagement strategies, making their essays more interactive and persuasive, whilst AI writing tended to be more neutral and positive rather than emotionally nuanced.
“We found that the essays written by real students consistently featured a rich array of engagement strategies,” Professor Hyland explains. “They were full of rhetorical questions, personal asides, and direct appeals to the reader—all techniques that enhance clarity, connection, and produce a strong argument.”
In contrast, the ChatGPT essays, whilst linguistically fluent, proved notably impersonal. They mimicked academic writing conventions but struggled to inject text with personal touch or demonstrate clear authorial stance. The AI consistently avoided questions and limited personal commentary, resulting in writing that was less engaging and persuasive overall.
The Broader Implications for Education
This research emerges at a pivotal moment for academic integrity and AI tools in education. Recent studies indicate that AI chatbots have thoroughly infiltrated academic publishing, with their use potentially extending even into peer-review processes, raising significant concerns among publication ethics experts.
The stakes are considerable: according to a Deloitte 2024 report, 79% of business leaders believe that generative AI will transform multiple organisations by 2027-2028, suggesting that understanding the limitations and capabilities of AI writing tools is becoming increasingly crucial.
The UEA study’s methodology was particularly rigorous, examining not just surface-level differences but the fundamental ways in which human and AI writers approach their relationship with readers. This focus on reader engagement represents a sophisticated understanding of what makes academic writing effective—it’s not merely about grammatical correctness or structural coherence, but about the writer’s ability to persuade, connect, and communicate complex ideas with personal investment.
Beyond Detection: Rethinking AI’s Role
Rather than dismissing AI tools entirely, the research suggests a more nuanced approach. Professor Hyland and his colleagues argue that tools like ChatGPT should be viewed as teaching aids rather than shortcuts. “When students come to school, college, or university, we’re not just teaching them how to write, we’re teaching them how to think—and that’s something no algorithm can replicate,” he emphasises.
This perspective aligns with emerging research on human-AI collaboration in academic writing, which suggests that the most effective approach may involve using AI tools to enhance rather than replace human writing capabilities.
The study also highlights the sophisticated nature of human engagement in academic writing. Students instinctively understand that effective communication requires more than merely presenting information—it demands establishing a relationship with the reader, acknowledging counterarguments, and positioning oneself within ongoing scholarly conversations.
The Statistical Learning Limitation
The research reveals a fundamental limitation in how current AI systems approach writing. ChatGPT’s statistical learning methods, whilst impressive in generating coherent text, prioritise consistency and coherence over conversational nuance and personal engagement. This reflects the nature of its training data and algorithmic approach, which excels at pattern recognition but struggles with the authentically human elements of communication.
This limitation has profound implications for how we understand the current state of AI writing tools. Whilst they can produce text that appears sophisticated on the surface, they lack the capacity for genuine perspective-taking and personal investment that characterises effective human communication.
Looking Forward: The Future of Academic Writing
The UEA study arrives at a crucial juncture in educational technology. As AI writing tools continue to proliferate, educators are grappling with questions about assessment, academic integrity, and the fundamental purposes of writing education.
The research suggests that rather than viewing AI as a threat to traditional academic writing, educators might consider it an opportunity to refocus on what makes human communication unique and valuable. The emphasis on engagement markers, personal perspective, and critical thinking skills may become even more important as these represent areas where human capabilities remain distinctly superior.
Practical Implications for Educators
The study offers practical insights for educators seeking to identify AI-generated content whilst supporting authentic student writing development. By focusing on engagement markers and personal voice, teachers can better distinguish between human and machine-generated text whilst encouraging students to develop their unique authorial voices.
Moreover, the research suggests that discussions about AI in academic writing should extend beyond detection and prohibition to encompass broader questions about the purposes of education. If the goal is to develop critical thinking and personal expression, then the human elements identified in this study become even more crucial to cultivate and assess.
The Enduring Value of Human Perspective
Ultimately, the University of East Anglia study reinforces something that many educators have long suspected: effective writing is about more than technical competence. It’s about the uniquely human capacity to engage with ideas personally, to take intellectual risks, and to communicate with genuine conviction and individual perspective.
As AI tools continue to evolve, this research provides valuable guidance for maintaining the integrity and purpose of academic writing education. By understanding what makes human writing distinctive—its personal engagement, rhetorical sophistication, and authentic voice—educators can better support students in developing these crucial capabilities whilst thoughtfully integrating AI tools where appropriate.
The study suggests that whilst AI may excel at producing grammatically correct and structurally sound text, the human touch—with all its imperfections, personality, and passionate engagement—remains irreplaceable in truly effective academic communication. In a time when technology has sidled up so closely to the human experience, it’s comforting to know that some of the essence of humanity is not replaceable.
We’d love your questions or comments on today’s topic!
For more articles like this one, click here.
Thought for the day:
“I believe the root of all evil is abuse of power.” Patricia Cornwell
