Human Connections in a Digital World

In a mid-level undergraduate history course, I asked students to reflect on what remains distinctly human about doing history alongside AI. Students encountered AI-generated interpretations of a piece of historical scholarship. They described the output as generic, overly confident, and less useful than reading, peer discussion, and sitting with uncertainty. Yet their own responses revealed a further tension: they did not always read AI-generated text to the same close, critical analysis they brought to assigned sources. My feedback sometimes mirrored their own critiques of AI, pointing to claims that were too general or insufficiently grounded in the text. Based on those responses, I redesigned a later assignment to ask not whether AI could interpret for students, but whether it could support narrower historical tasks without displacing human judgment. This session reflects on that trajectory in light of the American Historical Association’s emphasis on AI literacy and disciplinary judgment.