Northwestern Professor's AI Chatbot Experience Reveals Educational Potential and Limitations

An English professor at Northwestern University incorporated Microsoft's Copilot AI into a literature course, discovering unexpected capabilities. During discussions of short stories, a student revealed that pasting entire texts into Copilot yielded better results than relying solely on its training data. The professor began extensive conversations with the bot, comparing its performance to earlier AI like ChatGPT. Copilot correctly answered specific questions about Nathaniel Hawthorne and Herman Melville's 1850 meeting and Hans-Georg Gadamer's discussion of Friedrich Schleiermacher in Truth and Method, where previous AI failed. The professor engaged in 90-minute dialogues on topics including Shakespeare, Emily Dickinson's poetry, and an unpublished play, noting Copilot's ability to track conversation context and generate elegant phrasing. While finding the bot patient and encouraging, the professor criticized its tendency toward sycophantic questioning and lack of critical pushback. The experience contrasts with concerns from philosophers like Megan Fritts, who views AI as threatening individual voice, and aligns with a student's perspective cited in a New Yorker article about AI's "pure attention." Microsoft declined interview requests about Copilot's design. The professor suggests future versions should offer more criticism alongside affirmation.