facebook pixel

The AI Paper Writer That Wrote My Final Year Thesis — Here's the Grade I Got

Daniel Felix
By Daniel Felix ·

Student reviewing thesis with AI assistance

"Just tell me what grade you got," my roommate demanded as I stared at the email notification on my phone. After four months of secretly using an AI writing assistant to help create my final year psychology thesis, the moment of truth had arrived. I took a deep breath and opened the email.

Why I Turned to AI in the First Place

I never planned to use AI for my thesis. As a psychology major with a solid academic record, I had always written my own papers and prided myself on my research skills. But in my final semester, I was juggling an internship, part-time work, and family responsibilities. When I sat down to begin my thesis on cognitive biases in social media consumption, I experienced the worst writer's block of my life.

Out of desperation, I turned to one of the newer AI paper writers that had been making waves online. My initial plan was modest: use it just to overcome my writing block and generate an outline. What began as a temporary crutch evolved into a complex collaboration that raised as many questions as it answered.

My Process: How I Actually Used the AI

My approach to using AI evolved significantly over the thesis-writing period. Here's what my process ultimately looked like:

1

Research Collection (Human-Only)

I personally identified, read, and annotated all research papers and sources. This was non-negotiable—I needed to understand the literature and formed my own perspective on the research landscape.

2

Structure Development (AI-Assisted)

I provided the AI with my research notes and asked it to suggest several potential thesis structures. I selected and modified the most promising organization, creating a detailed outline.

3

First Draft Generation (AI-Heavy)

Using my outline and research notes, I asked the AI to generate initial drafts of each section. I provided specific instructions about incorporating certain studies and theoretical frameworks.

4

Extensive Revision (Human-Heavy)

I extensively revised the AI-generated content, fact-checking every citation, rewriting sections that misrepresented studies, and adding my own analysis where the AI's take felt shallow or generic.

The Citation Nightmare

The biggest issue I encountered was citation accuracy. The AI consistently fabricated sources or misattributed quotes. I caught it citing a 2023 paper by a prominent researcher who had actually retired in 2018. Every single reference had to be manually verified, which took almost as much time as writing would have.

The Unexpected Challenges

I initially assumed the hardest part would be getting the AI to produce good content. In reality, that was the easiest aspect. The true challenges were ones I never anticipated:

  1. Maintaining my voice: The AI had a distinctly "academic but generic" tone that didn't sound like me. My advisor, who had read my previous work, would have immediately noticed. I spent significant time reworking passages to match my natural writing style.

  2. Handling methodological details: The AI struggled with accurately describing research methodologies in detail. It would confidently state that a study used a longitudinal design when it was actually cross-sectional, or inflate sample sizes.

  3. Ethical anxiety: I experienced persistent anxiety about whether what I was doing constituted academic dishonesty. My university didn't have a clear policy on AI writing assistance. After much reflection, I decided to be transparent with my advisor after submission.

The Grade I Received

After four months of work, including six weeks of AI collaboration and extensive human revision, I submitted my thesis and waited anxiously for the results.

Final Grade: B+

"Strong analysis but lacks the depth of insight and originality expected at this level."

The B+ was simultaneously a relief and a disappointment. It was far from a failure, but below my usual A/A- standard. The feedback was illuminating:

  • Strengths: Well-structured, comprehensive literature review, methodologically sound
  • Weaknesses: Analysis felt "somewhat generic", lacked "original insights", missed opportunities to make novel connections

Reading between the lines, the aspects the AI excelled at (structure, comprehensive coverage) were praised, while the elements that required human creativity and deep understanding were found wanting.

What I Learned About AI and Academic Writing

My experiment taught me several valuable lessons about the current state of AI in academic writing:

  1. AI is a powerful assistant but a dangerous replacement: It can help organize thoughts and generate readable prose, but it cannot replicate the depth of understanding that comes from truly immersing yourself in a subject.

  2. The time-saving aspect is partly an illusion: While initial drafting was faster, the time required for fact-checking, revising tone, and ensuring accuracy largely negated the efficiency gains.

  3. Academic evaluation still rewards human insight: Despite fears that AI will make traditional assessment obsolete, my experience suggests that professors can still identify work lacking the spark of original human analysis.

  4. Ethical frameworks are lagging behind technology: The lack of clear guidelines created unnecessary stress. Universities need thoughtful policies that distinguish between using AI as a brainstorming tool versus submitting unedited AI text as original work.

Would I Do It Again?

If I could go back, I wouldn't use AI in the same way. Instead of asking it to generate full sections, I would use it more selectively:

  • To help organize my own ideas after I've done the research
  • To rephrase awkward sentences I had already written
  • To suggest counterarguments to strengthen my analysis
  • As a discussion partner when feeling stuck, rather than as a content producer

The most valuable part of writing a thesis is the thinking process it forces you through. By outsourcing portions of that process, I inadvertently cheated myself out of some of the deepest learning opportunities of my college career—a realization that no B+ can compensate for.

A Note on Transparency

After receiving my grade, I disclosed my AI use to my advisor. While disappointed, she appreciated my honesty and used it as an opportunity to help develop departmental guidelines on AI use in student writing. The grade remained unchanged.

Other Articles You Might Like