How Generative AI Can Support Professional Learning for Teachers

At the Harvard Graduate School of Education’s Creative Computing Lab, researchers partner with K–12 computing teachers across the country and around the world. Many of the teachers the lab supports help children develop early coding skills by using the popular programming platform, Scratch. With Scratch, kids program their own interactive media, like animations, games, and narratives, by snapping together digital blocks of code, just like building with Lego.

“When students are working on their Scratch projects, they will inevitably get stuck on something and not be able to figure out what's wrong. So they ask the teacher for help — and a lot of the work that teachers are doing is one-on-one with students to help them get unstuck,” explains Paulina Haduong, Ph.D.'23, a postdoctoral fellow in the lab.

Part of the teacher's role in supporting their students in creative and self-directed learning projects like Scratch is building up their ability to explore, discover, and problem-solve by themselves. However, "making this type of self-directed learning happen in practice" can be challenging for teachers, particularly when there are many students in their classrooms, according to Professor Karen Brennan, director of the Creative Computing Lab.

A new professional learning experience being launched this month by the lab tackles the problem head on. The free online tool, BlockTalk, uses generative artificial intelligence (AI) to provide teachers with a way to practice creative problem-solving with students’ Scratch projects.

On the website, teachers can work through a series of interactive conversational simulations. In the first of 10 “simulations,” for example, they are presented with a scenario called “Margaret's Dinosaur Dilemma.” After clicking on the simulation, a teacher would see that a young learner named Margaret wants to make her dinosaur roar, but her coding project isn’t working properly yet.

The BlockTalk tool “simulates a student. In the first simulation, Margaret is an AI student who states a problem that she’s having. Teachers can practice the interaction with the student, and they can do that multiple times,” explains Haduong.

After helping Margaret learn how to make her dinosaur roar, teachers can then try other conversations with other AI students about their Scratch projects, in simulations such as  “Reshma's Talking Sprites,” “Tim's Moon Landing,” “Jeannette's Dress Up Game,” and more, each of which addresses a different challenge that the researchers say real-life students have faced while learning to create Scratch programming projects.

Brennan and Haduong recently discussed how generative AI, as an emerging technology, can, through platforms like BlockTalk, support professional learning for teachers:

• Creates ways for teachers to practice interactions with students in low-pressure environments

  Teachers can practice interactions with “students” in lower-stakes situations online. “The hope is that BlockTalk helps teachers think about the interactions that they have in the classroom with students,” explains Haduong.

• Creates opportunities to collaborate and learn from others

  Teachers can review their simulated conversations with Margaret, Reshma, Tim, Jeannette, and other virtual students and then share them with other teachers. To be able to “see each other’s pedagogical practice and see each other’s pedagogical moves” is something that teachers who have already experimented with BlockTalk are “really excited about,” says Haduong.

• Creates invitations to focus on the learning

  Emerging technologies can create hope and excitement but also anxiety and concerns about the impact on learning, according to Brennan. She says the key is not to focus on what technology can or cannot do but to “stay centered on creativity, expression, self-direction and then ask the question: ‘How can technology help us to advance those aims?’ That’s where really magical things can happen.”