So What If Khanmigo Isn't Perfect. Is It Better Than the Alternative?
29 Feb 2024 · 1 min

0:55
In this brief commentary, Michael Horn reflects on a recent Wall Street Journal article that evaluated Khan Academy's AI tutoring system, Khanmigo, and found it frequently made basic math calculation errors. Horn frames these shortcomings as typical of early disruptive innovations, noting that while Khanmigo may not yet match top human tutors, most learners lack any tutoring support at all. He poses the central question: is Khanmigo an improvement over the alternative of having no tutor? Horn suggests that if Khanmigo can at least surpass the no‑tutor baseline, it has the potential to improve over time; if it cannot, its value is limited.










