Study finds no gender gap in AI code assistant effectiveness, but reveals cognitive benefits

A new study from arXiv has examined whether male and female programmers interact differently with AI code generation tools, and whether these differences affect how well they solve programming problems. The investigation sought to understand fairness and inclusion in human-AI collaboration as these tools become more central to development workflows.

The experiment#

Researchers recruited 39 programmers, balanced evenly by gender, who completed two medium-to-high difficulty coding tasks under different conditions. One group relied solely on AI code generation tools like Claude Code and GitHub Copilot, while the other had only internet access for reference and learning. The researchers used a counterbalanced design—varying task order and switching which tools people used first—to prevent biases from practice effects.

Throughout the study, scientists measured cognitive load using surveys and collected task performance data including completion time, code quality, and how people interacted with the AI tools. Screen recordings provided additional detail on interaction patterns.

What they found#

The headline result: no statistically significant gender differences emerged in cognitive load or task performance outcomes when using AI assistants versus internet-based workflows. This suggests that code generation tools level the playing field rather than advantage one group over another.

However, the study also documented that AI tools did offer broader benefits. Compared to internet-only workflows, code generation tools reduced both intrinsic cognitive load (the mental effort required to understand the task itself) and extraneous cognitive load (friction from the tool or process). Additionally, AI assistance improved the sophistication of the code people produced, even if completion time and basic correctness rates showed smaller gains.

What it means#

These findings hint that as development teams adopt AI coding assistants, they may be making programming more cognitively manageable for all developers—without inadvertently creating disparate experiences by gender. The research underscores that equitable access to these tools could support more inclusive and sustainable programming practices across diverse teams.