At our recent research retreat, we shared our experiences of AI tools for literature search, writing, and analysis. Since we are now reporting on AI’s usefulness as a ‘research assistant’, we also used ChatGPT to help draft this post— making it a practical experiment in itself!

🔍 Smarter Literature Searches: Elicit & Consensus

We’ve used these AI-powered search engines to discover academic papers.

• Elicit delivers structured insights, acting like a mini meta-analysis (and very recently even offering systematic reviews in paid versions).
• Consensus provides a broader mix of sources, including non-academic content.
• Our observation: They return different results for the same query, so cross-checking is valuable. We’re also still testing how Elicit performs when conducting systematic reviews.

🎨 Visualizing Ideas: Napkin

Napkin turns text into quick conceptual diagrams—perfect for research presentations or proposal sketches. We all enjoyed playing with it, and would recommend to just give its free version a go.

🕸️ Mapping Research: Research Rabbit

This tool creates citation networks, helping to identify key papers and connections. However, it leans on older publications, so it’s best for field overviews rather than cutting-edge research. You can tailor it a bit more when uploading your own collection and asking it to add suggestions for similar publications. However, we found the interface and visualization of the citation networks a little overwhelming at first.

🤖 ChatGPT & Other AI Assistants

ChatGPT remains a go-to tool, now with improved accuracy. Key uses include:

• Summarizing and analyzing uploaded papers.
• Generating feedback in a peer review style.
• Assisting with writing, though data privacy remains a concern.

Other notable tools include Perplexity, DeepSeek, and Aleph Alpha, although we have not used those extensively in our research

✍️ Writing Support: DeepL Write

Great for refining academic writing and maintaining consistent terminology, especially in multilingual work.

🎙️ Transcription Aids: Otter, Firefly & Dragon NaturallySpeaking

These AI-powered tools help transcribe interviews but still struggle with accents and poor audio quality. If interviewees agree, they can be a great time saver.

⚠️ Data Protection & Regulatory Uncertainty

A major challenge remains: university regulations on AI tools vary, creating uncertainty for researchers and students who collaborate across institutions. Here, we’ve used Napkin to outline our concerns and potential solutions when using AI tools. What do you think of the graphics?

Final Thoughts

AI isn’t a magic fix, but it can be a powerful research ally when used thoughtfully. There are so many social media accounts that introduce a variety of AI – we’ve decided to test a few of those in the coming months, so stay tuned for more discussions on AI and its usefulness for research.

What AI tools have you found useful in (sustainability) research? What camp are you on: team yay or nay AI? Share your experiences and thoughts in the comments.