Unless you’ve recently been hiding under a rock, you’re probably aware of the hype surrounding generative AI platforms and specifically ChatGPT. But you may not have considered how to effectively make use of this new tool as a researcher.
In this post we’ll explore how you can harness the full potential of ChatGPT to streamline your research process. From practical use cases and tips on mastering prompt engineering, to the ethics of AI assistance.
So, let’s dive in and discover how to use ChatGPT effectively as a researcher!
Sidenote – Honestly, I was slightly skeptical before writing this, however having tested it for the examples in this article I’m now impressed and plan to make use of it myself! It does have its limitations but I still see a value in making use of tools such as ChatGPT.
What is ChatGPT?
A Brief Overview of Generative AI
Before we discuss ChatGPT, I just want to briefly touch on the wider topic of generative AI. I constantly see headlines about generative AI in the news and on LinkedIn, and as a researcher it’s useful to be aware of what’s happening in this space.
Generative AI is artificial intelligence that is capable of creating content, whether that be text, audio, images, videos or other media (slideshows etc). Until the last few years, when AI tried to create things it usually wasn’t that great and there were no fears from creatives that their own work could ever be matched, but recently that’s all started to change.
Generative AI works by using existing content as a reference point and learning from it to make its own creations. But as it’s learning from work which humans created in the first place, that’s given rise to concerns around copyright and the urgent need for regulation. It also raises concerns around jobs, as seen with the strikes in America. Actors have raised concerns around unauthorised use of their images by AI and writers are worried about scripts being AI generated.
Things are moving quickly, but on a personal level I find it incredibly exciting that there are so many generative AI tools emerging that can allow anyone easily to realise their ideas, even if you have limited skills in a particular area. It’s becoming easier than ever to bring an idea to life.
I’d encourage all of you to keep an eye out as new generative AI tools are announced, to consider if they could have a role in assisting with your work. Beyond ChatGPT and other text-based tools, there are many generative tools for creating presentations, images, logos and more.
Keen to further your knowledge of generative AI?
If you’d like to learn more about generative AI you may find these free training courses useful:
Let me know in the comments if you’d be interested in a post covering the more broad topic of generative AI and the range of tools which you may find useful as a researcher.
An Intro to ChatGPT
So, where does ChatGPT fit in with all of this? Well ChatGPT is a conversational tool based on an advanced large language model designed to generate human-like text responses. In simple terms, you write a question and it will write a response.
It’s by no means the only large language model, but as of the date of writing it’s certainly the most well known. Another large language model you may have heard of is Bard from Google. For those who use the Bing search engine you may have also spotted BingGPT and this is built on top of the same technology as ChatGPT.
Each of these AI models has been trained on a vast body of text and they act as a reasoning engine. They take your question, work out what you’re asking and then formulate an answer to your query. It’s important to be aware that they can only provide responses based on the text they were trained on. Therefore, if that text contained any inaccuracies they will relate that information, as there’s no way for them to verify whether it’s true.
ChatGPT can write everything from poems and articles, to travel itineraries. It can also format information into tables.
You can currently access ChatGPT3.5 for free or you can pay to access ChatGPT4, which has an expanded knowledge of events beyond September 2021.
Prompt Engineering
To make the most of ChatGPT’s capabilities, it’s essential to master the art of prompt engineering.
A prompt is the text you provide to ChatGPT in order to request it to perform a task. This could be as simple as “Write 5 puns about cats”.

Not bad… especially considering it came up with them in less than four seconds!
But while simple prompts can be tempting to use, to increase the likelihood that ChatGPT can answer your question effectively it’s worth investing some time in crafting your prompts. This will greatly influence the chance that the responses meet your specific need.
There are loads of guides and courses popping up specifically for prompt engineering but here are four best practices for creating effective prompts:
- Be Clear and Specific: Make sure that your request is clear, with no room for ambiguity. The more specific you are, the better ChatGPT can understand and address your needs.
- Break Down Questions: Have a complex query? Consider breaking it down into simpler sub-questions. This helps ChatGPT to provide more accurate responses by addressing each component of your question individually.
- Provide Context: Adding relevant context to your prompt helps ChatGPT better understand your intentions and generate responses aligned with your requirements. This could involve explaining who you are (e.g. a researcher) and what you need the answers for (e.g. you’re planning an experiment).
- Experiment and Iterate: If ChatGPT doesn’t get it right on the first try, have another go. As it stores the memory of your conversation you can have a back and forth dialogue until you manage to secure the answers you’re looking for.
The makers of ChatGPT, OpenAI, have now released their own guide to Prompt Engineering which is worth checking out.
Example Prompts
Keeping the best practices above in mind, imagine you’re a researcher who wants some ideas for papers to read relating to their work in cartilage regeneration – this was the topic of my own PhD.
Take a look at the three prompts below. Which do you think would be most likely to give the most comprehensive answer?
A. What are the most important articles about cartilage?
B. What are the five most important research articles about the structure of articular cartilage?
C. I am a PhD student learning about the structure of articular cartilage. What are the five most important research articles published since 2010 for me to read? For each article please provide the reference for the article, a 100 word summary and state why each article is noteworthy.
The final prompt results in this detailed response, which provides the perfect launch point for further reading:

Remember, if you don’t like the result you can very easily iterate, making small tweaks to the wording of the prompt to get the desired outcome.
It’s also important to make use of the ability to have a conversation with ChatGPT, asking follow-on questions to drill down and gain a better understanding. We’ll go through an example of this shortly.
Now that we’ve discussed prompt engineering, let’s dive into some practical use cases for how you could use ChatGPT effectively during your research.
How to Use ChatGPT Effectively As a Researcher
Here are five ways you could leverage ChatGPT as a researcher:
Writing Assistance
Whether you’re drafting a thesis chapter, a research paper, or a conference abstract, ChatGPT could provide valuable assistance. Share your rough drafts or specific sections and seek input on sentence structure, paragraph coherence, grammar, or even suggestions for more effective arguments. ChatGPT could be a virtual writing partner, offering a fresh pair of eyes and helping you to refine your work.
This also applies to creating catchy titles for your research paper. Feed in your abstract and receive suggestions for the title. For example, I fed it an abstract of a paper I recently had accepted and asked for it to suggest a title for me.
Needless to say, I’m impressed!

Important: Just be mindful to not share any information that you don’t want stored by OpenAI.
Literature Review and Knowledge Expansion
Navigating through the vast expanse of academic literature can be overwhelming. However ChatGPT can serve as your personal research assistant, helping you to conduct literature reviews, summarising relevant papers and identifying key concepts within your field.
For example:

Since you can converse back and forth with ChatGPT, you can ask for clarification. For example, I asked why these articles in particular were important to read:

Although hallucinations are a very real thing you need to be aware of when using tools such as ChatGPT, I have found that it can often provide useful information when it can’t provide a solid answer:

ChatGPT can help to broaden your understanding, potentially enabling you to discover new perspectives and explore connections between various research studies. But always remember to critically evaluate and verify the information provided by ChatGPT and don’t use it as your only source for papers.
As mentioned earlier it’s also worth being aware that the free version currently only includes content up to September 2021, so you’ll have to look elsewhere (or pay) for content from the last few years.
Experiment Design
Designing rigorous experiments is integral to many research projects and it turns out this is another area where ChatGPT can come in useful.
ChatGPT can help you to refine your research methodology by suggesting alternative approaches, highlighting potential pitfalls, or offering recommendations based on existing studies. It can also provide an overview of considerations and required methods for conducting certain types of experiments.
For example:

Engaging with ChatGPT can help you to optimise your experiment design and ensure robust results.
However, it is still crucial to consult with your supervisor and colleagues to ensure the validity and reliability of your proposed research approach.
Presentation Preparation
Whether you’re getting ready for a conference presentation or your PhD viva, ChatGPT can help you to prepare effectively by suggesting common questions you may receive.
For example:

It can also share tips on how to communicate complex topics to audiences who may not share your in-depth understanding.
Skill Development
If you’re anything like me you’ll constantly be keen to acquire new skills and knowledge. ChatGPT can help by recommending relevant books, online courses, or research articles. You can also engage in discussions with ChatGPT to deepen your understanding of tricky concepts or seek explanations on topics outside of your area of expertise.
For example, one potential prompt for ChatGPT could be:

These examples are just some of the many ways for how to use ChatGPT effectively as a researcher. I’m sure you’ll come up with plenty of your own use cases too!
The Ethics of Using ChatGPT as a Researcher
Before using ChatGPT to assist your research it is important to consider the ethical implications of incorporating AI assistance.
Here are some key considerations to keep in mind:
- Acceptable usage: Before you use ChatGPT for a certain purpose, check that you’re okay to use it for that purpose. For instance, using it as an editing tool for drafts of your thesis might not be acceptable at your university. Also some journals have statements around acceptable use of tools such as ChatGPT. If you’re unsure, do your research and speak to your supervisor.
- Academic Integrity: While ChatGPT can be an invaluable tool, it is important to uphold academic integrity by clearly distinguishing between your original work and the contributions made by ChatGPT. Ensure that you give proper credit to ChatGPT or any other AI assistance utilised in your research when appropriate.
- Transparency and Bias: AI models like ChatGPT are trained on vast amounts of data which can be subject to inherent biases, these can carry through to ChatGPT’s responses. Be aware of potential biases and critically evaluate the information provided by ChatGPT. It is crucial to verify the accuracy of the generated content with reliable sources.
- Data Privacy: When interacting with ChatGPT, be mindful of the information you share. Avoid sharing sensitive research data or personally identifiable information, as the conversations may be stored or analysed.
- Responsible Use: ChatGPT can be a great tool to assist your research efforts, but do not rely solely on its outputs. Exercise critical thinking and maintain a balanced approach. Engage in discussions with colleagues and your supervisor to validate any insights gained.
Also be sure to check if your university has a policy around the use of generative AI. This will be the best guide of where it is and isn’t appropriate to make use of it. Here is an example of one from Imperial and another from Bristol.
Conclusion: How to Use ChatGPT Effectively
ChatGPT can be a helpful tool to assist you with your research. By mastering prompt engineering and following best practices, you can effectively utilise ChatGPT as a valuable resource for writing assistance, literature reviews, experiment design and skill development. However, ChatGPT should only be used to supplement your own expertise and critical thinking.
Incorporating AI assistance into your research process also comes with ethical considerations. It’s vital to uphold academic integrity by being transparent about AI contributions, critically evaluating the generated content and respecting data privacy. By using ChatGPT responsibly, you can leverage its power while ensuring the integrity and rigor of your research.
It’s an intriguing new addition to our research toolkits and I’m personally really excited to explore its potential!
Hopefully you now know how to use ChatGPT effectively!
Have you started experimenting with using ChatGPT or other generative AI tools in your work? Let me know in the comments below!
