What competencies do social scientists need to responsibly incorporate AI tools into their research practices?

Note: This post was link-posted from the Good Questions Review project by the Forum team, with the author's permission. The author may not see or respond to comments on this post. Summaries were auto-generated using GPT-4o. Any mistakes are our own.

In a new post for the Good Questions Review project (backed by Open Philanthropy), Paul Kellner examines what skills social scientists need to navigate generative AI. Drawing on recent reviews, he outlines key competencies—from critically assessing AI outputs and managing data ethics to developing better prompting and staying current with field norms.

The central tension is speed: Gen AI is evolving faster than academia typically moves. While the tools can boost creativity, analysis, and writing, they also bring risks around bias, transparency, and integrity. Kellner’s take is both grounded and forward-looking, urging researchers to engage with Gen AI thoughtfully and ethically. He also underscores the value of open-source infrastructure and institutional guidance—areas closely tied to EA’s work in field-building and AI governance.