AI tools for research: How to use AI tools smartly and responsibly
It feels like in 2024, there’s an AI solution for everything: right from grocery shopping to alerting you if your elderly mom living alone is likely to have a fall. And we can say the same about research: we’ve got AI tools to recruit our sample, clean our data, and even keep participants from dropping out. But these AI tools for research need to be used the right way; else, we’ll end up echoing The Guardian’s article that concluded “Forget artificial intelligence – in the brave new world of big data, it’s artificial idiocy we should be looking out for.”
Here are some popular AI tools for research, and examples of how to use them smartly and responsibly.
Academic Writing Tools like Paperpal
The right way:
- To generate ideas,
- To create a coherent article outline,
- To achieve the right academic tone,
- To make your writing more sophisticated,
- To translate complex academic content,
- To format your paper correctly,
- To get citation-backed, scientifically accurate support for your statements
- To check for plagiarism, and paraphrase and summarize text
The wrong way:
To spit out an entire paper with just 1-2 prompts, without verifying the accuracy and integrity of the output, and without disclosing to your target journal that you’ve used AI to write your paper
Image Creation Tools like Mind the Graph
The right way:
- To source graphics,
- To create clear and accurate data visualizations,
- To design infographics, graphical abstracts, posters, etc. to disseminate your research findings
The wrong way:
As a substitute for actual scientific output. If a journal peer reviewer asks you to add an image to the subsection describing your Western blot results, you cannot add a generic image of a Western blot. The reviewer wants to see your actual experimental data.
Literature search tools like R Discovery
The right way:
- To keep up with the latest trends in your field through a personalized reading feed,
- To access audio or translated versions of papers,
- To make sure you’re not overlooking preprints, conference proceedings, etc., in your literature search,
- To optimize and organize your literature search across multiple research projects
The wrong way:
- To blindly add citations into your paper without verifying that they are relevant
- To skip actually reading and critiquing the research the tool recommends
Clinical data workflows like Mendel.ai and Medidata
The right way:
- To speed up patient recruitment,
- To visualize complex data interactions in a chart review,
- To gather cohorts for simulation studies
The wrong way:
As a substitute for critical thinking and clinical decision-making
Journal finder tools like Editage’s Global Journal Database
The right way:
- To quick identify reputed journals that match the aims and scope of your paper,
- To verify a journal’s claims regarding indexing, impact factor, etc.,
- To rapidly locate journals that comply with your funder’s policies regarding open access
The wrong way:
To submit your manuscript to a suggested journal without verifying you meet the journal’s requirements regarding manuscript format, submission declarations, etc.
All-purpose large language models (LLMs) like ChatGPT
The right way:
- To create non-technical content like acknowledgements and dedications,
- To clean up basic language errors,
- To get explanations of terms/methods you don’t fully understanding (e.g., using the prompt “what are the steps for data cleaning in Python”)
The wrong way:
- To generate citations and references,
- To generate keywords according to Medical Subject Headings (MeSH),
- To translate technical content,
- To write an entire paper or entire sections of a paper without verifying that the content is accurate,
- To use without proper disclosure to your target journal
Bottom line
AI tools for research are here to stay, and researchers cannot afford to ignore these tools if they want to keep up in a competitive academic environment. But even as far back as 2006, Yudkowsky predicted the key problem with using AI tools: “By far the greatest danger of Artificial Intelligence is that people conclude too early that they understand it.”
When selecting any AI tool for research, you need to make sure that you’re allowed to use it, that you’re using it for the right purpose, you’ve kept your data safe (e.g., Paperpal has a strict data privacy policy and will not use your data for training), and that you’re still in control of the entire research and manuscript preparation process. AI tools can improve your research, but they can’t do your research for you.
Comments
You're looking to give wings to your academic career and publication journey. We like that!
Why don't we give you complete access! Create a free account and get unlimited access to all resources & a vibrant researcher community.
Subscribe to Manuscript Writing
Translate your research into a publication-worthy manuscript by understanding the nuances of academic writing. Subscribe and get curated reads that will help you write an excellent manuscript.
