SciSpace Academy
Tutorial

Research in the Age of AI: Using AI Without Losing the Researcher

Dr. Hélène NguemgaingEnglishResponsible AI

TL;DR

1-min read

AI now touches nearly every stage of research, from finding literature and developing ideas to analysing data and communicating results. In this tutorial, Dr. Hélène Nguemgaing, Assistant Clinical Professor at the University of Maryland, College Park, who leads the CRESA research stream and holds a Ph.D. in Sustainability Studies from West Virginia University, explains how to benefit from these tools without giving up critical thinking. It is aimed at researchers, faculty and student researchers who want rigorous, transparent and accountable work.

Key takeaways

  • AI can support literature search, idea development, data analysis and communication, but each use should be deliberate.
  • AI outputs need to be verified before they enter your research, by tracing claims and citations to sources that actually exist and say what was reported.
  • Disclosure of AI use should follow institutional and journal rules, and knowing when it is required avoids problems later.
  • The researcher stays accountable for the research questions, methods, evidence and conclusions, whatever tools were involved.
  • Protecting confidential data is part of responsible use, alongside accuracy and transparency.
  • A live SciSpace demonstration shows practical examples of where AI helps and where its output needs checking.
Apply it

To apply this, pick one task in your current project to support with SciSpace, verify every output against the original sources, and note how you used the tool so you can disclose it accurately.

How to do it

  1. Map where AI supports your research

    Identify where AI can help in literature search, idea development, data analysis and communication. This shows where to apply it deliberately.

  2. Try one task in SciSpace

    Run a practical task such as a literature search in SciSpace to see how AI assists. Hands-on use shows where its output helps and where it needs checking.

  3. Verify AI outputs before using them

    Check every AI output against the original sources before it enters your research. Verification is central to responsible AI use in research.

  4. Decide when to disclose AI use

    Determine when and how your AI use should be disclosed in research and publications. Transparency keeps your work trustworthy.

  5. Keep accountability for your conclusions

    Own your questions, methods, evidence and conclusions yourself, whatever AI contributed. You remain responsible for the work.

Watch the tutorial35 min

Responsible AI in Research: What to Delegate, Verify and DisclosePlay video: Responsible AI in Research: What to Delegate, Verify and Disclose

Chapters

  1. How do you use AI without losing the researcher?
  2. What is lost when AI writes your literature review?
  3. What should you delegate to AI, and what must you own?
  4. How can AI help you understand papers, not just summarize them?
  5. What does it really mean to verify AI output?
  6. How do you check AI-generated code and quantitative analysis?
  7. Can you share confidential or unpublished data with AI?
  8. How do you document prompts and model versions?
  9. How is AI used in peer review, and what must you disclose?
  10. Why do you remain responsible for AI-assisted findings?
  11. SciSpace demo: how effective is carbon pricing at reducing emissions?
  12. How do you find relevant research papers and authors?
  13. Do the papers cited together actually agree?
  14. Why can't you compare findings reported in different units?
  15. What conclusion does the evidence support?
  16. How do you verify an AI summary before citing a paper?
  17. Could you defend your AI-assisted work in a PhD defense?

Try this on SciSpace

Frequently asked questions

What does responsible AI use in research mean?

It means using AI tools in ways that keep research accurate, transparent and accountable: checking outputs against original sources, protecting confidential data, disclosing AI use according to institutional and journal rules, and making sure the researcher, not the tool, makes the key decisions about questions, methods and conclusions.

How do I verify information produced by AI?

Trace every factual claim and citation back to its original source and confirm that the source exists and says what the AI reported. Re-run or check calculations independently, compare results across sources, and be cautious with confident statements that lack evidence. Treat AI output as a draft that needs review, not as a finished answer.

Related tutorials