Ask HN: Is using AI tooling for a PhD literature review dishonest?

I'm a professor of computer science, and I use AI to conduct literature reviews all the time. It's actually one of the handful of things that AI really excels at. So really, I think I would be more inclined to teach students how to do use AI effectively rather than discourage it. I know I would have loved to have it when I was working on my PhD, and I haven't authored a paper in the past 2 years that did not include help from LLMs.

I tend to use Notebook LM to gather and analyze papers, then I synthesize some notes, outline the ideas and prompt it to give citations to the outline. Then I take that and write my literature review narrative, which I then hand over to claud for proofreading.

End to end, the process is much faster than back when I did it manually. Just like typesetting a paper in latex is faster than the hot metal process of 50 years ago. The bottom line for me is this: You should always use every tool at your disposal when doing real work, and research is real work.

Someone against AI will tell you yes, someone for AI will tell you no. The only thing I can really say is that saying you have ADHD so you should have a reprieve from the normal rules is something that I don't agree with.

While your dashboard sounds fancy, this part raises issues:

> I run ChatGPT Pro to collect all relevant papers

Any literature review must be reproducible. If you can't say exactly what queries you ran against exactly what databases, you'll get into trouble. Whether or not that's the way things should be is irrelevant: it's the way things are.

You should ask your supervisor if your approach is okay. If necessary, ask it from a theoretical perspective: "would it be okay if I were to....?" If your supervisor is unavailable then seek advice from their colleagues.

Since you mention ADHD, you're likely to be strongly motivated by novelty. Don't spend time building a dashboard that you could spend on writing your thesis. If you're not getting support from your university, get it now. It might not help, but it's a signal to the university that you're engaging with the system.

The verification pipeline is the most valuable part of your workflow. Most people who use AI for literature reviews skip exactly that step — they trust the output and move on.

What you're describing is closer to building a testing harness than "using AI to write." You're asserting claims, checking them against source PDFs, and reviewing manually. That's more rigorous than most manual lit reviews where people skim abstracts and cite papers they half-read.

Document the pipeline as methodology in your dissertation. That turns a potential misconduct question into a contribution.

> The idea of the dashboard is the following: I run the Codex through a web chat to identify the relevant quotes — relevant for my dissertation topic — and how they are relevant, it combines them into a number of claims connected with each quote with a link. And then I review each quote and each claim manually and tick the boxes.

I’d be most concerned about this component of your process, tbh. IIUC, you’re not just using the LLM to identify relevant papers, i.e. a fancy search engine. You’re also extracting specific statements, divorced from their context in a given paper, and using these to make claims for your research.

Even if you validate that the quotes are actually present in the papers, are you also reading the full papers to ensure you understand the overall results of the paper and what the quotes mean in that context? Or are you just identifying hopefully-relevant snippets and combining them?

Yes and no. The first thing to understand is that in academia, knowledge is the work. You are being trained to absorb existing knowledge, hypothesise new knowledge and test if it is valid.

LLMs are a useful tool if you want it to generate text. But in the context of research, this is quite dangerous. Think of a calculator that spits out the wrong answer 10% of the time, would you trust it to use in an exam? How about 5%? 1%? 0.1%? The business of research is the business of factual knowledge. Every piece of information should and is expected to be scrutinized. That's why dishonesty is severely looked down upon (falsifying data / plagiarism etc.)

I would say your use case is not dishonest, but I would also like you to think from the perspective of the university. How would they know if their students are using it honestly like you did? How can they, with their limited resource, make sure that research integrity is upheld in the face of automated hallucinations?

At the end of the day, the question is not what if using AI is dishonest, it's about being able to walk into an antagonistic panel and defend your claim that you understand the knowledge of your field (without live AI help). If you can do that and also make sure that the contents are not hallucinated, then I don't see why not.

Everyone dislikes pedantically verifying references. However, if you cut corners here then will you also cut corners pedantically verifying research results?

Beyond references, the point of the literature review is to ensure you have read the literature and understand it well enough to accurately summarize it. If you present a literature review, it's likely assumed you did all of this. So at the very least you should be upfront about how an LLM assisted you.

You cannot copy others' work and claim it is your own. Thus, you cannot copy ChatGPT's work and claim it is your own. There is a qualitative difference between having an LLM generate text and having a program spell- and grammar check text. Since you are not going to highlight which passages in your article ChatGPT wrote for you and instead intend to pass it of as your own creative work it is dishonest. Very dishonest. If caught you will get in trouble and may be kicked out of your academic programme.

Think about it this way: 70 years ago, would a physicist be considered a cheater for using a calculator to solve complex differential equations in their daily work? People tend to frame the moral dilemmas of new technology through the lens of everyday human tasks, and I think that's just a prejudice.

No, I don't think it is dishonest.

At the same time I would recommend, document your methodology explicitly in the dissertation, describe the verification pipeline, and make it clear what you reviewed manually versus what was automated. That transparency converts "dishonest?" into "methodologically rigorous."

Here is the thing, academic policy is NOT really about honesty. It is about trust. Universities cannot distinguish your workflow from someone who prompted GPT to write their lit review wholesale.

More than the ethical distinction, I believe the rule around AI usage is blunt because enforcement is pretty hard.

I don’t know if it is dishonest. What I do know is that it will only save you time if you have a very specific and testable need. Otherwise it will appear to save time and produce something that you won’t be proud of.

I don't think what you are doing is dishonest. But my opinion hardly matters.

My advice is to talk to your dissertation committee chair to understand whether they think it is dishonest. Furthermore, read your university's AI usage policies. If they don't consider what you are doing a permissible use of AI, no amount of assurance on HN or any online forum is gonna help you.

Not dishonest if you verify everything and understand it deeply but you should be transparent about your AI use since many universities care more about disclosure than the method itself.