Khanmigo does not directly answer students’ questions, but instead begins with questions of its own, such as whether the student has any ideas on how to find an answer. It then guides the student toward a solution, step by step, with suggestions and encouragement.
Despite Khan’s expansive 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 get stuck or stuck in a cognitive rut, he says, “we want to help students get out of a rut.”
About 100,000 students and teachers tested Khanmigo last academic year at schools across the country, helping to spot any delusions the bot was having and providing tons of conversations between students and bots for DiCerbo and his team to analyze.
“We look for things like summarizing, hinting and encouraging,” he explains.
It’s not yet clear how much Khanmigo has closed the AI engagement gap. Khan Academy plans to release some summary data on student-robot interactions later this summer, according to DiCerbo. Plans to have outside researchers evaluate the tutor’s impact on learning will take longer.
AI feedback works both ways
Since 2021, nonprofit Saga Education has also been experimenting with AI feedback to help tutors better engage and motivate students. In collaboration with researchers at the University of Memphis and the University of Colorado, Saga team driver in 2023 They fed transcripts of their math tutoring sessions into an AI model trained to recognize when the tutor was asking students to explain their reasoning, refine their answers, or engage in deeper discussion. The AI analyzed how often each tutor followed these steps.
By tracking some 2,300 tutoring sessions over several weeks, they found that tutors whose coaches used AI feedback peppered their sessions with significantly more of these prompts to encourage student engagement.
While Saga is considering having AI deliver some feedback directly to tutors, it is doing so cautiously because, according to Brent Milne, vice president of product research and development at Saga Education, “having a human coach in the loop is really valuable to us.”
Experts expect AI’s role in education to grow and for its interactions to continue to seem more and more human. Earlier this year, Open AI and startup Hume AI separately thrown out An “emotionally intelligent” AI that analyzes tone of voice and facial expressions to infer the user’s mood and respond with calibrated “empathy.” However, even emotionally intelligent AI will likely fail to engage students, according to Brown University computer science professor Michael Littman, who is also the director of the National Science Foundation’s division of information and intelligent systems.
No matter how human the conversation, he says, 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 their classes. In turn, the students will never really care about the robot and what it thinks. A June Study In the journal Learning and Instruction, it was found that AI can already provide decent feedback on student essays. What is not clear is whether student writers will put in the care and effort, rather than delegate the task to a robot, if AI becomes the primary audience for their work.
“The human relationship component of learning is incredibly valuable,” Littman says, “and when we simply remove humans from the equation, something is lost.”
This story about AI Tutors was produced by The Hechinger Report, an independent nonprofit news organization focused on inequality and innovation in education. Subscribe to the Hechinger Report Hechinger Newsletter.