AI, History, and the Fight for Source Truth
Thu Jan 29 2026
AI makes it dangerously easy to stop at one confident answer. For education publishers and learning platforms, that’s a product risk: teachers need sources they can trust, students need research support (not shortcuts), and credibility doesn’t survive black-box answers.
In this episode, Jan van der Crabben of World History Encyclopedia explains how they built the platform on one core idea: history is a web of connections, not isolated timelines.
Jan also shares how their History AI evolved through teacher feedback, why shorter answers and visible citations matter, how academic journals help reduce hallucinations when proprietary content is limited, and why trustworthy publishing is at risk as AI becomes the default interface to knowledge.
More
AI makes it dangerously easy to stop at one confident answer. For education publishers and learning platforms, that’s a product risk: teachers need sources they can trust, students need research support (not shortcuts), and credibility doesn’t survive black-box answers. In this episode, Jan van der Crabben of World History Encyclopedia explains how they built the platform on one core idea: history is a web of connections, not isolated timelines. Jan also shares how their History AI evolved through teacher feedback, why shorter answers and visible citations matter, how academic journals help reduce hallucinations when proprietary content is limited, and why trustworthy publishing is at risk as AI becomes the default interface to knowledge.