AI for research and scholarly publishing

Overview
Key points and resources
Disclosure
How can AI help you? The Good
Issues with AI for scholarly work (The Bad)

---

Overview: two basic types of AI in scholarly publishing

• Generative, e.g. ChatGPT
• Predicative AI, machine learning algorithms “learn from and make predictions or decisions based on data, have found numerous applications in publishing. They can predict article impact or citation rates, recommend relevant articles to readers, identify potential reviewers for manuscripts, and detect patterns in research trends,” (Lo).
  

  ---

  Key points and resources:

  • Know publisher policies regarding use of generative AI in writing; approach use of AI with a critical lens. AI can be a huge timesaver for tasks including ike data analysis (predictive AI) but should be used judiciously for writing-related tasks (cross-check generative AI results).
  • Keep up with Gen AI tools of all types Generative AI Product Tracker – Ithaka S+R.
  • Overview articles:
    Bergstrom, Tracy, and Dylan Ruediger. A Third Transformation? Generative AI and Scholarly Publishing. Ithaka S+R, 2024. https://doi.org/10.18665/sr.321519.
    Kousha, Kayvan, and Mike Thelwall. “Artificial Intelligence to Support Publishing and Peer Review: A Summary and Review.” Learned Publishing 37, no. 1 (2024): 4–12. https://doi.org/10.1002/leap.1570 [includes various journal finder tools]Gibney, Elizabeth. “What Are the Best AI Tools for Research? Nature’s Guide.” Nature, February 17, 2025, d41586-025-00437–0. https://doi.org/10.1038/d41586-025-00437-0.

  ---

  Disclosure

  • Use of AI for writing isn’t disclosed by many authors despite journal policies.
  • Willingness to disclose use of AI for writing is influenced by discipline (Brainard)
  • How do authors acknowledge ChatGPT? Chiefly for writing! (Kousha). However, we should take these findings with a grain of salt since many authors do not disclose use.

    

  Check journal policies re. disclosure of use of AI; policies are not helpful to authors and may be vague.

  ---

  How can AI help you? (The Good)

• Brainstorming Chat GPT or other generative AI potentially useful for brainstorming for subtopics, outlining, and expanding on your idea.

• Your literature review:

  • Some library search tools currently or eventually will include an AI tool (library tools don’t provide for natural language searching)
  • Watch out for hallucinated citations!
    • OpenAI’s GPT-5 is supposed to be much better in that regard
    • Math and law especially problematic (Gibney)
  • Can only be assessed by comparison to traditional modes of inquiry. Play around with these two alternatives to Google Scholar: Semantic Scholar vs Elicit.org vs Google Scholar—all different
    • Semantic Scholar seems like a souped-up Google Scholar
      • Create an account for all features including alerting
      • Filtering by discipline powerful
      • Speed up your reading at least for items in arXiv via Semantic Reader
      • good to have it in case Google Scholar goes away
      • More of a search engine than not
    • Elicit = AI summaries and search tools
      • Go to Elicit (https://elicit.org)
      • Input a research question
      • review the AI-generated summaries.
      • tool’s ability to find and filter relevant papers.
      • Claims it can do systematic lit reviews
      • Able to extract data from papers!
      • Free version limited to 50 papers?
  • NotebookLM (Google)
    • Seems to mostly search the web
    • Has a notebook feature to save and upload sources
  • Scite
    • not free
    • works with publishers
  • Other search tools: Perplexity, Ai2Paper Finder similar to ChatGPT, unclear how good they are, what they search
  • Litmaps and ResearchRabbit—visualization of search results that show connections between articles
  • Reminder: LLMs are NOT search engines

• Data analysis

  • Provalis (not free)
  • RapidMiner (not free)
  • Tableau has AI built in?
  • Julius AI
  • Note programming languages and skill level needed
  • learn more, “The AI Revolution in Data Science (Feat. Paul Groth).”

• Generative AI in writing

  • • Overview from Clemson U.
    • Grammarly
    • Other products: Paperpal, Writefull and Curie
    • Very nice for feedback and polishing for small bits of your writing {I have done that}

  Five levels of writing assistance by LLMs, with example prompts, (Lin).

  Level	Prompt
  (1 ) Basic editing, such as checking spelling and grammar, or suggesting synonyms.	?Check the spelling and grammar in this paragraph, and suggest synonyms for any repetitive words.?
  (2) Structural editing, such as paraphrasing, translating, or improving the structure of the text, or its flow or coherence.	?Paraphrase this lengthy sentence to improve its clarity and flow, and translate it to French.?
  (3) Creating derivative content, such as summarizing, creating titles and abstracts, rewriting or generating analogies.	?Summarize this document and create a short, catchy title for a journal submission.?
  (4) Creating new content, such as completing, continuing or expanding text, or brainstorming ideas.	?Continue the text to explain the key question being addressed. Show why it is important, drawing parallels or analogies where you see fit.?
  (5) Evaluation or feedback, such as assessing the quality of the writing or finding weaknesses in it.	?Review this introduction and highlight any logical gaps or areas that need further development.?

• Other great uses of AI

  • (both types) can check manuscripts for errors in coding, mathematics, statistics, or logic + fact checking [Black Spatula Project is exploring effectiveness]
  • Abstract and title creation
    • Copilot, chatgpt, claude.ai, gemini
  • Translation as well as adaption for other audiences
• Learn how to craft good prompts!
  

  ---

  Issues with AI for scholarly work (The Bad)

• Think critically about AI, not a panacea. “[we must] reckon with what these systems can and cannot do. They summarize but do not listen, reproduce but do not interpret, and excel at coherence but falter at contradiction,” (Burzlaff).
• Potential detection and rejection
  • Know the journal’s policies about use of AI-generated text!
  • “tortured phrases” (language created by generative AI) are pretty apparent to others. Similar to students plagiarizing. May be a red flag to editors about academic integrity. Tools can detect use of generative AI in writing, (Naddaf).
• AI reinforces and replicates biases in scholarly publishing, e.g. biases against authors from certain countries, gender bias
• Authors and publishers are concerned about unauthorized, uncompensated training of their copyrighted content for LLMs

plus!

• AI for screening and peer review is problematic, ineffective and very commonly used by publishers.
  • Many editorial organizations have guidelines about ethical use of generative AI and other AI tools, (Syring, Alison, and Gibbs) (Hodgkinson).
  • Some journals allow use of AI in peer review, NIH does not allow, (Li et al.)
  • Not uncommon in computer science, (Zou).
  • Potentially disruptive to knowledge (epistemic) communities
• Although generative AI can highlight retracted and other discredited articles, it is not accurate (Thelwall et al.)
  • It may boost discovery of questionable journal articles because they are not behind publisher paywalls
• However, extensive use of ChatGPT to write papers is rare (Desaire)

sources

Brainard, J. (2025). Far more authors use AI than admit it. Science, 389(6766), 1168–1169. https://doi.org/10.1126/science.aec3552

Burzlaff, J. (2025). Fragments, not prompts: Five principles for writing history in the age of AI. Rethinking History, 1–18. https://doi.org/10.1080/13642529.2025.2546174

Desaire, H., Isom, M., & Hua, D. (2024). Almost Nobody Is Using ChatGPT to Write Academic Science Papers (Yet). Big Data and Cognitive Computing, 8(10), 133. https://doi.org/10.3390/bdcc8100133

Gibney, E. (2025). Can researchers stop AI making up citations? Nature, 645(8081), 569–570. https://doi.org/10.1038/d41586-025-02853-8

Hodgkinson, M. (2025, September 16). Help or hindrance? Peer review in the age of AI. DOAJ Blog. https://blog.doaj.org/2025/09/16/help-or-hindrance-peer-review-in-the-age-of-ai/

Kousha, K. (2024). How is ChatGPT acknowledged in academic publications? Scientometrics, 129(12), 7959–7969. https://doi.org/10.1007/s11192-024-05193-y

Li, Z.-Q., Xu, H.-L., Cao, H.-J., Liu, Z.-L., Fei, Y.-T., & Liu, J.-P. (2024). Use of artificial intelligence in peer review among top 100 medical journals. JAMA Network Open, 7(12), e2448609–e2448609.

Lin, Z. (2024). Techniques for supercharging academic writing with generative AI. Nature Biomedical Engineering, 9(4), 426–431. https://doi.org/10.1038/s41551-024-01185-8

Lo, L. S. (2025). Generative AI and Open Access Publishing: A New Economic Paradigm. Library Trends, 73(3), 160–176.

Naddaf, M. (2025). AI tool detects LLM-generated text in research papers and peer reviews. Nature. https://doi.org/10.1038/d41586-025-02936-6

Newsletter, T. D. S. E. C. (2025, February 21). The AI Revolution in Data Science (feat. Paul Groth) [Substack newsletter]. The Data Science Education Community Newsletter. https://datascienceeducation.substack.com/p/the-ai-revolution-in-data-science

Syring, A., & Bueno Gibbs, G. (2025, September 17). Artificial Intelligence and Peer Review in Scholarly Publishing. Feeding the Elephant: A Forum for Scholarly Communications. https://networks.h-net.org/group/discussions/20125690/artificial-intelligence-and-peer-review-scholarly-publishing

Thelwall, M., Lehtisaari, M., Katsirea, I., Holmberg, K., & Zheng, E. (2025). Does ChatGPT Ignore Article Retractions and Other Reliability Concerns? Learned Publishing, 38(4), e2018. https://doi.org/10.1002/leap.2018

Zou, J. (2024). ChatGPT is transforming peer review—How can we use it responsibly? Nature, 635(8037), 10–10. https://doi.org/10.1038/d41586-024-03588-8