A few years ago, I used to teach a session on “writing a literature review” for a general “methods” course that was offered to incoming graduate students in various humanities and social sciences field.
I always found teaching this somewhat nonsensical because 1) if you read scholarship, you can see what a literature review is: it’s right there in virtually everything you read.
2) There are plenty of videos on YouTube that tell you exactly what to do.
So, I used to cover the basics, and then talk about related issues, like what predatory journals are and how to identify them. But I would also direct students to a YouTube video on writing a literature review, and there were several good ones that were out there.
Then ChatGPT arrived on the scene. Fortunately, I stopped teaching that session not long after that point, and the reason I say “fortunately” is because in the past three years, since ChatGPT first emerged, a lot has changed in the “YouTube advice for graduate students” world.
To show this, I am going to link to one YouTuber’s videos below, a gentleman by the name of Andy Stapleton. Andy has 333K subscribers and has made videos for the past 7 years or so. As such, he was making videos long before the emergence of AI, but he has perhaps been among the most active in showing people how to incorporate AI into their research and writing.
I have absolutely nothing again Andy, and I would even congratulate him for being a pioneer in seeing what this technology can do. However, as will become evident from viewing the topics of his videos over time (and I encourage you to explore his YouTube channel and watch his videos), this technology can do a lot of what graduate (and undergraduate) students and scholars have traditionally done on their own.
In what follows, I will link to some of Andy’s videos over time, to give a sense of how the information that he provides to graduate students has changed as AI and AI tools have arrived on the scene.
Again, I’m making no value judgements about any of this. I am doing this simply because I see this as a reality that I still find many people are not yet taking seriously.
If any of this surprises you, let me be the first to welcome you to the brave new world of higher education in the AI age.
August 11, 2020 (255,776 views): How to write a literature review – my simple 5 step process!
June 6, 2023 (251,124 views): 3 Unbelievable AI Technologies to Automate Your Literature Review
Nov 21, 2023 (470,311 views): How To Write An Exceptional Literature Review With AI [NEXT LEVEL Tactics]
Dec 14, 2023 (53,575 views): How to use AI to find *ALL* the literature for your research | A blended approach
Feb 13, 2024 (127,852 views): What Is A Literature Review? Ditch Old Methods for Cutting-Edge Tech!
Aug 15, 2024 (72,946 views): AI Just Made Literature Review a Piece of Cake
Oct 1, 2024 (68,956 views): This AI Tool Does Literature Reviews in SECONDS (100x faster than you)
Jan 28, 2025 (76,010 views): This AI Tool Wrote My Literature Review – and It’s a Game-Changer
Jun 19, 2025 (40,682 views): One Prompt = Full Literature Review (Why Every PhD Needs This Tool)
Jul 15, 2025 (20,998 views): Elicit AI Can Now Do Your Entire Systematic Review (In Minutes)
Aug 14, 2025 (63,735 views): This AI Tool Crafts an Entire Research Paper From a Few Notes
I assume that in another two years or so, my field — political science — will start seeing PhD dissertations that are ~ 95 percent AI generated, and the small proportion of “traditional” writing by the doctoral student will only be icing on the cake. Maybe the entire genre of the dissertation will disappear, except for narrow academic subfields where generating new data remains the accepted research norm. A dissertation that is really nothing but a 500-page literature review is just pointless.
It’s hard to see the point of having AI do this. I want to survey the literature myself, understand the scope and range of its arguments, their evidentiary basis, theoretical grounding, and lacunae. A machine can possibly do these things, but not -for me-. Since I will view all these things through my prior idiosyncratic formation. And only then can I generate my own arguments and positions in relation to that literature. Having AI do a literature review for a research topic is like having the janitor or my grandmother do such a review. Neither have any background or bildung. Just probabilistic token associations. Nothing resembling propositional thought or logical or material implication and inference. There is no I in AI.
Hello Haydon,
At the university where I most recently worked, there was huge pressure to get research grants from the university.
What did people do with those research grants? They hired research assistants to 1) write the literature review, 2) collect the data, and 3) transcribe the data (and in some cases they probably did more than that).
I was present at numerous seminars where profs presented on their research, only to see them have to defer to their research assistants when questions came up because they themselves had no idea how to answer the questions.
This is already the norm at tons of universities around the world, and these are precisely the conditions where technological advances take over. Technological advances do not replace successful/solid practices. They replace failing practices. Research and writing is already a failing practice at a huge percentage of the universities in the world, and that provides a perfect space for AI to enter and take over.
You may have reviewed article manuscripts from scholars from a certain country that we are both familiar with. One of the weakest parts of those manuscripts has been the literature review section. Their content/data might be workable, but they don’t know the literature to place it in. AI solves that problem for them. Again, a place where something is failing is precisely the place where AI will take over.
I think that in the US, the biggest threat from AI is in the realm of teaching (a failing world in many ways and cases). So, people might keep doing their own literature reviews (although I think whatever younger people still come up through the system will use AI), but they’ll find their departments getting reduced in size as “pedagogical advances” from AI replace humans.