In brief: Only one in five social scientists uses autonomous coding agents, despite their potential to revolutionize research processes. Clear disparities emerge by gender and institution—a sign of growing digital inequalities in academia.
A survey of 1,260 social scientists shows: while 81 percent use AI chatbots, only 20 percent have integrated autonomous coding agents into their research. Adoption is highly uneven—researchers with typically male names use these tools twice as frequently as female researchers.
Automated coding agents like Claude Code could fundamentally transform the social sciences. They can independently write, execute, and interpret analyses—tasks previously considered indispensable for human involvement.
However, the representative survey from February and March 2026 reveals substantial disparities in usage. Researchers at top-tier universities deploy the tools 40 percent more frequently than their colleagues elsewhere. Users of coding agents publish more papers and grant proposals, though this could also reflect differences among early adopters.
While scientists are optimistic that AI helps in writing publication-ready papers, they express concerns about the overall impact on their field. The potential consequences are complex: the technology could accelerate research and reduce costs, but it could also exacerbate inequalities and strain the scientific publishing landscape.
Source: www.anthropic.com