In a crowded fourth-grade classroom in Chicago, a new kind of tutor is shaping how children learn about empathy, conflict, and problem-solving. These robots aren't programmed to act like friendly classmates with invented emotions and backstories. Instead, they speak plainly, without pretense or fiction, and the results will catch educators' attention across the country.
The research behind it, led by PhD student Lauren Wright and overseen by Assistant Professor Sarah Sebo at the University of Chicago's Department of Computer Science , came together thanks to robust partnerships: Chicago Public Schools (CPS) provided access to classrooms and teachers, and Kiljoong Kim at Chapin Hall built crucial connections that made this cross-institution project possible.
"With this work, we wanted to create a team that would be able to uniquely design and study technology, informed by best practices in SEL education, with the input of principals, administrators, teachers, and students in Chicago Public Schools," said Sebo. "We started not with a specific robot prototype, but by observing SEL instruction in CPS classrooms and talking with teachers about their experiences with SEL, and THEN starting to think about how robots might be able to supplement the amazing work teachers are already doing in schools."
This research was presented at the 2026 ACM/IEEE International Conference on Human-Robot Interaction (HRI) held in Edinburgh, Scotland, the top venue for human-robot interaction research, where the team not only shared their findings with the broader academic community, but also learned they had won the Best Paper Award—an honor that recognizes the most impactful and innovative research at the conference.
For teachers in CPS, SEL lessons usually mean whole-class activities delivered once a week. In practice, however, many students tune out, and overstretched teachers would love more one-on-one opportunities.Teachers interviewed in the study expressed concern that group SEL lessons rarely reach every child. This perspective, along with careful classroom observation and interviews, drove the research team to look for solutions.
Wright's team spotted the gap and asked whether robots could supplement teachers and provide individualized instruction where group lessons fall short. Did it actually matter if those robots 'acted' human?
In the experiment, fifty-two students participated. One group learned SEL from robots with fictional, emotion-laden dialogue. Another worked with robots that spoke only in factual terms, openly acknowledging they had no feelings or friends. The third group received their regular curriculum with no robot involvement. By leveraging Second Step's curriculum, made possible through connections with Chapin Hall and Kiljoong Kim, the research team ensured the lessons reflected real classroom needs. Robots adapted group lesson plans into personalized conversations, while teachers continued focusing on the rest of their students.
Both robot groups showed students improved in their mastery of SEL concepts compared to peers who only had classroom instruction. Yet, the factual robots, in their straightforward honesty, often encouraged deeper engagement with lesson vocabulary and problem-solving language. These findings challenge conventional wisdom.
"Giving robots fictional personalities with the intent to make them more engaging is a common approach to educational robots, one which feels especially relevant for teaching SEL," expressed Wright. "However, in our research study, we found that the robot's fictional emotions and experiences may have distracted or made students feel less comfortable using lesson language. These findings challenge us to reconsider our assumptions when designing robot behaviors – just because an approach is common doesn't mean it will always lead to the best outcomes."
Honest Robots, Authentic Impact
As society becomes more concerned about children forming unhealthy attachments to AI, the Chicago team's results provide timely guidance. Demonstrating that factual robots can perform as well or better without mimicking emotions points the way to a safer classroom technology.
The central message of the study is clear. Robots are powerful supplements, extending teachers' capabilities and freeing up attention for students who need more support. But they do not replace the human element in teaching.
"We firmly believe that human teachers are the most important element in elementary education," said Sebo. "As we all experienced during the pandemic, replacing in-person educational experiences with technology-mediated ones can be disastrous. Our work does not seek to replace human teachers, but instead, aims to create robot tools that extend a teacher's reach, giving the ability to provide children with one-on-one attention without pulling them away from the rest of the class."
As this school year winds down, Chicago's classroom experiment stands as proof of what partnership-driven innovation can achieve in education. The findings invite other districts to rethink how technology can responsibly supplement teachers and ensure every child receives meaningful, individualized support.
