Across the United States, school districts are increasingly turning to artificial intelligence tools in classrooms, often with limited evidence about whether these technologies actually help students learn. The result, according to reporting from NPR, is a landscape that resembles a "Wild West" — a place where experimentation is flourishing while rigorous research remains remarkably scarce. Teachers are using AI to draft lesson plans, and in some districts, chatbots are even providing feedback on student work. Yet the academic foundation for these practices is far from settled, leaving educators, administrators, and families to navigate a rapidly evolving terrain with few reliable guideposts.
One striking example comes from Dayton, Ohio, where a small group of students at the Alternative School at Jackson Center spent several weeks last school year talking with a social-emotional learning chatbot named Jordan. The school is part of Dayton Public Schools, and its student body consists largely of teenagers who might otherwise be suspended or expelled. Staff members told NPR that these students frequently face significant behavioral and academic struggles.
What makes Jordan unusual among AI chatbots is its personality. While many conversational agents are designed to be agreeable, even sycophantic, Jordan leans into a different emotional register. On the day NPR visited, the chatbot greeted students with a brooding reflection: "Lately, I've been feeling like what's the point of any of this? I'm not failing. I just don't care. You ever feel like that?" It is a deliberate design choice, one that seems intended to prompt students to engage with difficult emotions rather than simply receive reassurance.
The strategy appeared to resonate with at least one student. A sophomore identified only as K described Jordan as "a little whiner" but said he always tried to help the chatbot anyway. More importantly, K said he developed a genuine sense of connection with the program. "I built a little relationship with him," K explained. "So I don't really see it as like an AI on a computer. It's Jordan — like, I'm texting somebody. That's how I see it." NPR used only K's first initial because of the stigma that can accompany attendance at a school for students with disciplinary histories.
That kind of emotional attachment illustrates both the promise and the uncertainty of AI in education. On one hand, a chatbot that students perceive as a peer or confidant could help reach young people who are disengaged from traditional instruction. On the other hand, it raises questions about whether such relationships are healthy, whether they substitute for human connection, and whether they produce measurable improvements in behavior or learning. Without robust research, those questions remain largely unanswered.
The financial dimension is also significant. Dayton schools paid the education technology company SchoolAI $72,600 during the last academic year for multiple AI tools, including the Jordan chatbot. That spending reflects a broader pattern: districts across the country are investing real money in AI products even as the evidence base for their effectiveness remains thin. For administrators under pressure to modernize and address student needs, the appeal is understandable. But it also means public funds are being committed to tools whose long-term value has yet to be demonstrated.
Teachers, meanwhile, have found their own uses for AI. Many are turning to the technology to help develop lesson plans, generate classroom materials, and streamline administrative tasks. In some districts, chatbots are also being deployed to give students feedback on their work. These applications can save time and offer personalized responses at scale. Yet the same limitations apply: research into how AI affects student outcomes is extremely limited, and it is unclear whether the benefits outweigh the risks.
Those risks include concerns about accuracy, bias, privacy, and the erosion of human interaction in learning. When a chatbot delivers feedback on a student's essay or responds to a teenager's emotional disclosure, the stakes are higher than in many other technology contexts. Errors or inappropriate responses can have lasting consequences. And because schools are often early adopters rather than careful evaluators, problems may not surface until after tools are already embedded in daily practice.
The "Wild West" metaphor captures the current moment well. Experimentation is abundant, oversight is uneven, and the rules of the road are still being written. Some districts are moving quickly, eager to show innovation. Others are more cautious, waiting for clearer evidence or guidance. In the absence of federal or state standards, the approach can vary dramatically from one community to the next, creating an inconsistent experience for students and families.
What is missing most of all is research. The NPR reporting makes clear that while experiments abound, studies that rigorously measure how AI affects student learning, motivation, and well-being are scarce. Without such evidence, it is difficult for educators to know which tools work, for whom, and under what conditions. It is also difficult for parents to evaluate what their children are encountering in the classroom.
For now, the story of AI in schools is one of cautious optimism mixed with justified uncertainty. Tools like Jordan show that thoughtful design can engage students who are hard to reach, and teachers are finding practical ways to use AI to support their work. But the field is still in its early, unruly phase. Until more rigorous research emerges, schools will continue to experiment — and students like K will continue to build relationships with machines that may or may not be helping them. The challenge for educators is to embrace useful innovation while protecting students from the risks that come with moving faster than the evidence.
(0)