A team led by a Harvard scientist found that young students learn more from educational television shows when they have the opportunity to interact with artificial intelligence-powered versions of the show's characters.
“Children spend a lot of time watching television, and high-quality educational programming is an important source of learning,” says Ying Xu, a professor at the Harvard Graduate School of Education and lead author of the study that describes this research. “We also know that they learn best when an attractive adult sits next to them, asks them questions and talks to them.”
The problem is that caregiver availability does not always align with children's screen time, leading to potential loss of learning opportunities. Thinking about this, Xu wondered if ai chatbot technology could replace the role of the caregiver.
“What if we could integrate these technologies into children's television shows, allowing them to interact with the characters and learn in an engaging way?” Xu asks.
How was this study carried out?
To test their theory, Xu and his colleagues partnered with PBS Kids to conduct a study with more than 200 children ages 4 to 7. The students were divided into three groups of 80 students. All students were shown the PBS KIDS science program, Elinor wonders whywhich is aimed at preschool and early primary school children.
The show was combined with different levels of interaction. A group has just watched the program. Another group watched the show but also interacted with an ai-assisted version of Elinor, who is a curious cartoon rabbit, who encouraged children to answer questions and offered advice if they got it wrong. The last group of students watched the entire show with a section in which Elinor asked the viewers questions and gave a generic answer after a pause. “This is a popular approach that many TV shows use to create a sense of interactivity,” Xu says.
The students were then asked a series of questions to assess what they had learned.
What did this study find?
“We found that children who watched the interactive version performed better, answering most of the questions correctly, followed by those who watched the pseudo-interactive version, and lastly, children who watched the streamed version,” Xu says. “Our findings suggest a possible scalable mechanism to support children's learning through television. “We also confirmed that engaging in conversations with children helps improve their learning from television, even when these interactions are with characters rather than a human companion.”
However, there were limitations. Although statistically significant, the difference between the three groups was smaller than expected, Xu says. “Given that this was a very short study, we might anticipate larger effects when children watch multiple videos over a longer period of time,” he says. “This is exactly the type of study we plan to do in a couple of months.”
What's next for this type of research?
Beyond studying how the impact of interactive ai characters influences young learners over time, Xu is working to bring this concept to more natural learning environments.
“We are conducting research to understand how children might interact and learn from these interactive videos in their homes, rather than in the lab, as this is a more natural environment where most of their media consumption occurs,” Xu says. . “We are also exploring whether the benefits could extend to other STEM domains, such as computational thinking.”
Additionally, Xu wants to learn more about the real-world applications of ai-assisted lessons. “We are particularly interested in examining how interaction with ai could help children not only understand educational content but also improve their ability to apply what they learn to solve real-world problems,” Xu says.
