AI Can’t Replace Teaching, But It Can Make Teaching Better
Khanmigo does not answer students’ questions directly, but starts with questions of his own, such as asking if students have any ideas about how to find the answer. He then guides students step by step to find the solution, with suggestions and encouragement.
Despite Khan’s grand vision of “amazing” personal tutors for every student on the planet, DiCerbo assigns Khanmigo a more limited teaching role. When students are working independently on a skill or concept but are stuck or stuck in a cognitive rut, she says, “we want to help students get unstuck.”
About 100,000 students and teachers tested Khanmigo over the past school year in schools across the country, helping flag any hallucinations the bot caused and providing tons of conversations between students and the bot for DiCerbo and her team to analyze.
“We look for things like summaries, suggestions, and encouragement,” she explains.
The extent to which Khanmigo bridges the gap in AI interactions remains to be seen. Khan Academy plans to release some summary data on student interactions with the bot later this summer, DiCerbo said. Plans for third-party researchers to evaluate the tutor’s impact on learning will take longer.
AI Response Works Both Ways
Since 2021, the nonprofit Saga Education has also been testing AI feedback to help tutors better engage and motivate students. Working with researchers from the University of Memphis and the University of Colorado, Saga Team Pilot in 2023 fed transcripts of their math tutoring sessions into an AI model trained to recognize when tutors were prompting students to explain their reasoning, refine their answers, or initiate further discussion. The AI analyzed how often each tutor took these steps.
Tracking about 2,300 tutoring sessions over several weeks, they found that tutors whose coaches used AI feedback incorporated more reminders into the sessions to encourage student engagement.
While Saga is looking to use AI to provide live feedback to tutors, they’re still treading carefully because, according to Brent Milne, vice president of product research and development at Saga Education, “having a real coach is really valuable to us.”
Experts predict that the role of AI in education will increase and its interactions will continue to appear more and more human. Earlier this year, OpenAI and separate startup Hume AI has been released “Emotionally intelligent” AI is capable of analyzing voice and facial expressions to infer a user’s mood and respond with calibrated “empathy.” But even emotionally intelligent AI may fall short of its goal of engaging students, according to Brown University computer science professor Michael Littman, who is also director of the National Science Foundation’s division of information and intelligent systems.
No matter how human-like the conversation, he said, students understand at a fundamental level that the AI doesn’t really care about them, what they have to say in their papers, or whether they pass or fail the class. In turn, students will never really care about the bot and what it thinks. A June study in the journal Learning and Instruction found that AI was able to provide good feedback on student essays. What’s unclear is whether student writers would be interested and put in the effort, rather than handing the task off to a bot, if AI became the primary subject of their work.
“There is incredible value in the human relationship component of learning, and when you take people out of the equation, that’s lost,” says Littman.
This story is about AI Tutor Produced by The Hechinger Report, an independent, nonprofit news organization focused on inequality and innovation in education. Sign up to Hechinger Newsletter.
