10 Mistakes to avoid in Academic Writing


Tool Review: Consensus 2026 for Lit Reviews & Writing Phase

Dear Reader,

Academics who use AI for academic writing typically fall into two camps: Those who try to do too much at once, generating whole paragraphs, and those who only occasionally ask for small corrections.

One is ethically questionable and produces poor-quality output, while the other leaves a lot of potential on the table. Both require a lot of manual editing and don't make you all that much more productive.

Having tried both ends of the spectrum, I realised that the problem was me, not AI.

The solution I found for using AI productively is preparation.

If you provide what to write in a specific format, AI can save days of time when writing academic work. I call it the atomic sentence method, and if you have already signed up for the upcoming academic writing webinar, you will learn all about it shortly.

This week, I wrote down all the mistakes I made during the writing phase of my PhD and the last 4 papers. I hope you can avoid making the same ones.

Summary: The mistakes described in this article are: Missing or meaningless context, using AI for parts of the paper which are not suitable for AI, wrong prompts, wrong expectations, a lack of persistence to get good output, and a lack of precision. I also outline how to fix each of these mistakes.

Webinar: Writing With AI 2026 • Apr 11th

If you have high standards for academic writing, using AI is very challenging. The devil is in the detail, and the first bout of euphoria quickly fades when you realise that the output often lacks depth, finesse and individuality.

After 3+ years of experiment I found ways to solve it.

And the techniques presented in this webinar made me 2-3x more productive, even allowing me to complete my PhD requirements ahead of schedule.

Don't miss the webinar (or recording if you can't make it).

Bonus 1: My AI assistants (CustomGPTs)

For a year or so, I have been optimising a suite of AI-assistants for various tasks, which I will share with webinar participants.

My favourite of these AI assistants extracts ideas from papers into a format you can use to write your literature review:

The trick is that you don't take notes in just any format, but a specific one (the atomic sentences mentioned earlier) that is optimised for easier academic writing.

What did others say about the webinar in 2025?

The previous webinar in 2025 was quite a success, and I hope some of you who joined previously put the ideas to good use. Here are some voices from last year:

Here is what we'll cover:

  • The "Atomic Statement Method" to control what AI writes
  • Building AI assistants specifically for academic writing
  • Teaching an AI to write in your voice instead of a generic tone
  • Fact-checking your writing with AI
  • Skim papers with AI to deepen specific arguments
  • Improve clarity and flow of manuscript drafts
  • Understand which parts of the paper are suitable for AI writing
  • An iterative writing improvement workflow which avoids plagiarism
  • Understand when not to use AI for writing
  • Ethical considerations of AI writing (journals and universities)

Check the link below for more information & the date of the Webinar:

Wishing you a wonderful weekend,


Ilya Shabanov
The Effortless Academic

This is the weekly Effortless Academic Newsletter. It consists of an in-depth tutorial and additional events, promotions, or relevant information for the AI-curious modern academic. I strive to consistently provide value with every email. If this is not relevant to you, you can always unsubscribe from everything and will never hear from me again. If you find the email somehow inappropriate, please reply to me and let me know what you didn't like.

The Effortless Academic

Literature Review Tools, Note-Taking Strategies and AI tutorials for the modern academic. Publish more with less effort and supercharge your career.

Read more from The Effortless Academic

Paperpal (Sponsor) For researchers juggling multiple tools to check grammar, plagiarism, and AI detection, Paperpal now brings everything together in one place. With the new full paper check, you can upload your draft and run all three checks at once, helping you catch issues early and move to submission with greater confidence.» Try Paperpal Document Check Discount Codes:ILYAPRIME30 (30% off) Avoid overwhelm & procrastination through multi-lens project management Dear Reader, Most academics...

Conceptual Literature Review Dear Reader, A literature review starts with reading. But what is reading other than unpacking packaged information into a set of ideas that link together in meaningful ways in your head? What if we could start the process with the mindmap? This is what I call a conceptual literature review. Given a few papers, it uses AI to extract concepts and claims and create an interlinked information network. The output is an Obsidian vault with 100s of connections and...

Consensus (Sponsor) Consensus is one of the most advanced AI literature search tools utilising both the reference network and AI semantic similarity to find relevant papers and summarise their findings. I've used Consensus almost daily while writing my PhD. » Try Consensus Discount Codes:effortlessacademic (25% off) The most minimal knowledge management system for beginners (+ video walkthrough) Dear Reader, I recently introduced a colleague to Obsidian and have pitched it to our lab many...