Amid all of the hype about how genAI could be used in research, we hope this paper provides some grounded insight about how the scholars contributing to the sociological literature are *actually* using this technology.

Finally, we found that, despite low levels of trust (previous figures and lefthand figure here) and mixed responses about genAI's potential net positive effect (righthand figure here), there is still faith that genAI will continue to improve (middle figure here)

Another respondent connected this to work/life balance:

However, if the faster output it helps to generate just increases expectations for that much more output, then it may increase the number of studies done but won’t be of much personal benefit to stressed and overworked researchers/students.

The concern about misinformation also extended to a concern about critical thinking, something that surfaced through open ended responses like this:

"At this stage, GenAI and AI tools pose a threat to the integrity of knowledge that is factual, thoughtful, reflective, nuanced, and critical."

However, respondents were generally worried about genAI and the spread of misinformation, control by large tech companies, and costs related to the environment. They were less concerned about economic and computational costs, however.

Among scholars using genAI at least weekly, there is some agreement that genAI can save time and is already incorporated into tools they typically use. People are also just curious! They also felt less outside forces, reporting low levels of pressure from the field/dept/collaborators to use genAI

First, we find that a little over 1/3 of respondents report using generative AI at least once a week, a number in line with analogous studies from other disciplines. There weren't clear differences between computational and non-computational scholars, either (ended up being a recurring theme).

Note about the figures:

Yellow = computational sample
Purple = general sample

We sampled authors of papers that appeared in 50 sociology journals over the past five years, so it might be more accurate to say our study examines generative AI and sociology rather than just "sociologists."

We also pay special attention to scholars using computational methods in their research

Work done with @dustinstoltz.com @oms279.bsky.social and Marshall Taylor

Thanks bro!

Cool paper! You might find my work on this topic helpful/useful as well:

link.springer.com/article/10.1...

journals.sagepub.com/doi/abs/10.1...

arxiv.org/abs/2509.08199

arxiv.org/abs/2503.20062

Here is the preprint! arxiv.org/abs/2509.08199