Poor quality manuscripts are a waste of time for peer-reviewers like me, because there is no hope they will get published. Before the pandemic in 2020, I used to review manuscripts often. However, when I noticed I was getting a lot of poor quality manuscripts to review in the wake of the pandemic, I just stopped. I needed a break.
According to this peer-reviewed article:
Scientists are increasingly overwhelmed by the volume of articles being published. The total number of articles indexed in Scopus and Web of Science has grown exponentially in recent years; in 2022 the article total was ∼47% higher than in 2016, which has outpaced the limited growth—if any—in the number of practicing scientists.
These authors call it “the strain on scientific publishing”. A post on the International Science Counsel blog calls it “overpublication”, and discusses the kerfuffle with Wiley (a legitimate scientific publisher) acquiring Hindawi (considered by many a predatory publisher) in the context of this overpublication problem. I have my own criticisms of the Wiley/Hindawi connection – especially when what I consider “garbage journals” from Hindawi and MDPI are indexed.
But these articles are talking about published manuscripts. I’m talking about receiving poor quality manuscripts for review. Theoretically, the editor of the journal should not be sending poor quality manuscripts out for review.
After taking a few years off from reviewing papers, I restarted in 2026, and the results are not looking good. I made a graphic to show you.
As per the graphic, so far, I have accepted 12 manuscripts for review – 11 from the publisher Springer Nature. Of those, 7 were not worthy of review and should not have even been forwarded to me by the reviewer. By “not worthy of review”, I mean all you had to do was look at it and realize it was not a scientific paper. Before 2026, I would not have received such poorly written manuscripts for review. What’s going on?
What Happens When You Review Poor Quality Manuscripts?
Earlier this year, my colleagues and I published a paper about the difference between using AI and statistical models in scientific research. This may have helped me get on journal lists as a potential reviewer of papers on the topic of AI, because this year, of the 12 I’ve reviewed, as per the graphic above, 5 have had something to do with AI. Most – but not all – of the others relate to topics I’ve published on previously, such as medical education.
So the topics of the manuscripts I received were legitimate. But why were the manuscripts so crappy? The graphic below summarizes the current status of all 12 manuscripts as of this writing.
As you can see, of the 5 that were either “worthy of review” or “poorly done but worthy of review”, 2 have been accepted for publication, 2 are still under review, and 1 was rejected. Remember, a rejected manuscript represents a waste of time to both the reviewers (who are unpaid) and the journal (which has limited resources).
Now let’s look at the 7 that were “not worthy of review” in the first place, and here is the current status of those: 4 were formally rejected; 2 that I reviewed earlier in the year (May and June) never were revised and resubmitted by the authors (probably because the reviews were so bad), and 1 is still under review. I expect that last one will get rejected also.
What Kind of Tripe Am I Receiving to Review?
…and when I say “tripe”, I mean it. You don’t even have to read the manuscript to be able to tell there is something fundamentally wrong with it. Here are examples of what I mean:
- The manuscript is filled with non-printing characters.
- The manuscript is missing tables, or has only one or two tables that are overloaded with information, do not make sense, or both.
- The manuscript contains bullet-points
- The manuscript contains many headings and subheadings with only a sentence or a very small paragraph after them.
- The manuscript contains words that seem to have been randomly printed in boldface throughout the text.
…and this is before even reading it. This is just from eyeballing it. Remember, when you agree to review and article, you are only shown the abstract. If I could see the whole manuscript first, I would definitely not waste my time reviewing manuscripts like this.
Okay, so you might think: What if you read it? Maybe the authors don’t know the difference between an essay in college and a peer-reviewed article. Maybe it’s actually well-written.
Unfortunately, if you read it, it just gets worse. It usually doesn’t say anything. For example:
- Papers are cited, but not described (especially reports of studies – what are the findings???).
- Papers lack citations where they belong (e.g., after claims, and when summarizing results in reviews).
- Descriptions of concepts (in both the introduction and the discussion) are very general and do not really explain the concepts, or connect them directly to the topic at hand.
Okay, so I’m not naïve. I realize these papers were likely mostly if not completely written using some large language model (LLM) like ChatGPT or Gemini. But my question is…
Why Are Editors Passing on Poorly Written Manuscripts for Review?
If I can just look at the manuscript and – without even reading it – tell it is tripe, why is the editor of the journal even sending it out for review? I posed this question to one of my co-authors on the paper I mentioned earlier, who has expertise in AI tools.
He explained that a lot of scientific journals are automating their process of filtering in articles for review using AI tools!
…which is disgusting. This is a complete failure of the system of science. Not only that, he pointed me to this NPR article about how legitimate writers (like me) are having their writing being erroneously flagged as written by AI!
So not only are we wasting time and resources of reviewers and journals, we are also possibly rejecting good, solid writing before it even gets reviewed, and accepting complete tripe for review. What could we do about it? Here’s my recommendation…
Modify the COPE Guidance on “Artificial Intelligence in Decision-Making”
COPE stands for “Committee on Publication Ethics”. It is a non-profit organization run out of the United Kingdom that makes recommendations on how to run journals ethically. Many publishers require all their journals to follow COPE guidance, which can be found on their web page. You can access their guidance by becoming a member or else signing up for a free guest account and logging in.
COPE already has a guidance on “Artificial Intelligence in Decision-Making” (last updated in 2021). The guidance is over 10 pages, but you can download my one-page summary here. When it comes to editorial decisions, the guidance says:
At this point in time in the development of AI and the evolution of mindsets, we recommend that if an editorial decision is made by AI that provides a final outcome for an article, such as acceptance or rejection, the decision should directly involve an editor. The decision cannot be made by an AI tool alone.
However, it doesn’t say anything about sending manuscripts for review. I feel like that needs to be added to this guidance – that choosing whether manuscripts are sent for peer-review or not “should directly involve an editor” and “cannot be made by an AI tool alone”.
Read all of our data science blog posts!
Poor quality manuscripts are being passed onto reviewers, thus wasting our time. My post discusses the 12 manuscripts I’ve reviewed so far this year in 2026.

