Author says LLMs cannot fully automate EDA

The author argues that exploratory data analysis (EDA) cannot be fully automated by large language models because human intent, context, and organic hypothesis generation remain essential. While AI tools can handle repetitive coding tasks, they frequently make poor analytical decisions and lack the ability to intuitively catch subtle data quirks and edge cases. Consequently, human data scientists are still required to manually explore datasets to make meaningful discoveries. Open article →