AI tools now part of everyday life for SciLifeLab researchers
A recent survey among SciLifeLab/DDLS Fellows and their research groups reveals that large language model (LLM) tools, such as ChatGPT and Copilot, have become a regular part of SciLifeLab researchers’ daily work. The tools are primarily used for writing and coding, and many respondents expressed a desire for continued support in the form of premium access, training, and policy guidance.
The survey, conducted in December 2024, collected responses from 69 individuals within the SciLifeLab/DDLS Fellows research groups. Nearly 40 percent reported using LLM tools several times per day, and another 17 percent used them several times per week. Only 4 percent had never tried any AI tools. The most common applications were academic writing, such as research articles, grant proposals, and presentations, as well as scripting for data analysis.
“GitHub Copilot reduced my time spent on troubleshooting simple issues and allowed me to focus on the analysis of the data,” one researcher noted.
ChatGPT was by far the most popular tool, followed by Microsoft Copilot. Some respondents also mentioned Perplexity, Gemini, Claude, and Le Chat (Mistral), though these were used much less frequently. Many SciLifeLab researchers expressed a strong interest in further support from SciLifeLab. The most common request was access to premium tools, such as ChatGPT Plus or Claude Pro. Others called for more training opportunities and clearer guidelines, something that Ola Spjuth (Uppsala University), SciLifeLab AI Lead, is currently working on.
“Researchers around the world are adopting LLMs in their research, especially younger scientists and PhD students. This survey shows that SciLifeLab researchers are following the same trend and highlighting the importance of strategic support and engagement from SciLifeLab moving forward”, says SciLifeLab AI lead, Ola Spjuth (UU)
While the December survey focused on researchers within DDLS Fellows research groups, earlier data suggest that infrastructure and technical staff may use LLM tools differently. A broader survey conducted during the SciLifeLab Facility Forum in June 2024, with 352 participants, showed that only 10 percent of infrastructure staff used LLMs daily, and 21 percent had never tried one. Many in that group expressed greater interest in using AI to support users, for example through onboarding tools, FAQs, or automated assistance.
“We want custom support bots that help users find the right facility or person. I think that could save us hours,” one participant wrote.
This discrepancy may partly be explained by the fact that the first survey was conducted almost six months earlier. Still, the contrast highlights how expectations and applications of AI can vary depending on professional role.
The December survey makes clear that LLM tools have become a natural part of research at SciLifeLab. With continued support in the form of tool access, training, and policy guidance, both SciLifeLab researchers and infrastructure staff stand to benefit even further from these evolving technologies.
