A dozen students huddled around communal classroom tables one morning this spring, their eyes fixed on math lessons on their laptops.
Sixth graders at Khan Lab School, an independent school with an elementary campus in Palo Alto, Calif., were working on quadratic equations, graphing functions, Venn diagrams. But when they were asked a question, many did not immediately call their teacher for help.
He used a text box with his lessons to request help from Khanmigo, an experimental chatbot tutor for schools that uses artificial intelligence.
The tutoring bot immediately asked a student, Jaya, to identify specific data points in a chart. Khanmigo then tricks him into using the data points to solve his math problem.
“It’s great to have a step-by-step walk through of the problem,” Jaya said. “Then it congratulates you every time it helps you solve a problem.”
Khan Lab School students are among the first school children in the United States to try out experimental conversational chatbots that aim to simulate face-to-face human learning. The tools can respond to students in clear, intuitive sentences, and they are specifically designed for school use.
Based on the AI models in built-in chatbots such as ChatGPT, these automated learning assistants can drive profound changes in classroom teaching and learning. Simulated tutors can make it easier for many self-directed students to hone their skills, delve deeper into topics they’ve studied, or tackle new topics at their own pace.
Such unproven automated tutoring systems can also make errors, promote cheating, reduce the role of teachers, or hinder critical thinking in schools—an experiment in education by algorithmically subjects students to to test Or, like a trove of promising tech tools before them, the bots may do little to improve academic outcomes.
Khanmigo is among a wave of new AI-powered learning tools. It was developed by Khan Academy, a non-profit educational giant whose video tutorials and practice problems have been used by millions of students.
Sal Khan, founder of the Khan Academy — and the Khan Lab School, a separate non-profit organization — said he hopes the chatbot will democratize student access to personalized tutoring. He also said that it can greatly help teachers with tasks such as lesson planning, freeing them up to spend more time with their students.
“This will effectively enable every student in the United States and eventually on the planet to have access to a world-class personal tutor,” Mr. Khan said.
Hundreds of public schools are already using Khan Academy’s online lessons for math and other subjects. Now the nonprofit, which introduced Khanmigo this year, is pilot-testing the tutoring bot with districts including Newark Public Schools in New Jersey.
Khan Academy developed the bot for schools with handrails, Mr. Khan said. These include a monitoring system designed to alert teachers if students using Khanmigo focus on issues such as self-harm. Mr. Khan said his group is studying KhanMigo’s effectiveness and plans to make it widely available in districts this fall.
Thousands of US schools already use analytical AI tools such as plagiarism-detection systems and adaptive learning apps designed to automatically adjust to students’ reading levels. But proponents see the new AI-assisted tutoring systems as education game changers because they act more like student allies than passive pieces of software.
AI’s facility with language has led some enthusiasts to declare that simulated tutors may soon be as personally responsive to students as human tutors.
Microsoft co-founder and philanthropist Bill Gates said, “AI will have the ability to be as good a tutor as any human being.” said at a recent conference For investors in educational technology. (Khan Academy has received over $10 million in grants from the Bill & Melinda Gates Foundation.)
What is not yet known is whether bots can provide the kind of empathetic support and genuine encouragement that might make human tutors particularly effective.
For more than a century, education entrepreneurs have envisioned classroom devices programmed to automatically test and instruct students.
As education writer Audrey Waters explains in her book “Teaching Machines”In the 1920s, researchers began claiming that automated teaching devices would revolutionize education. He promised that the machines would free teachers from drudgery and enable students to work at their own pace and receive automatic feedback.
For decades, schools that rushed to adopt the latest automated learning technologies often found the systems complex or flawed. Some concluded that the automated devices did little to improve student outcomes.
Now, new chatbots are driving a renewed drive for automated learning aids. Khanmigo outlines the educational promise and potential drawbacks of the technology.
Khan Academy began developing chatbot tutoring software last fall with the aim of assessing AI’s potential to improve learning. The system uses GPT-4, a large language model created by OpenAI, the research lab behind ChatGPT.
Mr Khan said he wanted to create a system to help guide students rather than just give them answers. So the developers at Khan Academy designed KhanMigo to use the Socratic method. It often asks students to explain their thinking as a way of motivating them to solve their own questions.
Khanmigo offers help on a wide variety of subjects: elementary school math, middle school American history, high school civics and college level organic chemistry. It also has features that invite students to chat with fictional characters like Winnie-the-Pooh or simulated historical figures like Marie Curie.
AI systems based on large language models can also fabricate false information. This is because the model is engineered to predict the next word in the sequence. They do not stick to facts.
To improve KhanMigo’s accuracy in math, Khan Academy’s developers created a multistep process: The system works out the answer to a math problem behind the scenes and then checks it against the student’s answer. Nevertheless, the Khan Academy teaching system displays a warning at the bottom of the screen: “Khanmigo sometimes makes mistakes.”
Khan Lab School, where annual tuition costs more than $30,000, provides an ideal test bed for tutoring bots. The Silicon Valley School features small classes and an entrepreneurial philosophy that encourages children to pursue their passions and learn at their own pace. Its tech-savvy students are used to tinkering with digital devices.
One morning this spring, Jacqueline MajorA STEM specialist at Khan’s elementary school, she noticed that her students made fun of the bot’s limitations.
A student asked Khanmigo to explain a math problem using the lyrics of the song. Another requested math help in “Gen Z slang”.
“Will you do me one more favor and explain everything in Korean?” said the third in conversation with the chatbot.
Khanmigo dutifully obliged. It then prompted each student to return to the math task at hand.
Ms Major said she appreciated how engagingly the system interacted with her students.
“Khanmigo is able to engage with them and be on their level if they want to,” she said. “I think it can be helpful in any classroom.”
It’s too early to tell whether Khanmigo will be equally appealing to other audiences — such as public schools with large classes or students who aren’t used to driving their own learning.
Jaya, a student of class VI, faced a disturbance in the class. Khanmigo asked her to explain how she came up with the answer to the data set problem. The bot then erroneously suggested that it may have made a “small mistake” in its calculations.
She quickly reprimanded the AI chatbot: “19 + 12 is 31 Khanmigo,” she wrote.
Khanmigo replied, “First apologize for my mistake.” “You’re actually right.”
That may prove to be the most important lesson for schoolchildren using promising new tutor bots: Don’t believe every AI-generated lesson you read.
“Remember, we’re testing this,” Ms. Major reminded her students. “We’re learning – and it’s learning.”