Across American classrooms, artificial intelligence is arriving faster than anyone can measure its effects. In one striking example from last school year, a handful of students at the Alternative School at Jackson Center in Dayton, Ohio, spent several weeks talking with a social-emotional learning chatbot named Jordan. The tool was not designed to flatter students or tell them what they wanted to hear. Instead, Jordan was built to seem moody and withdrawn — a digital personality with something closer to teenage angst than corporate cheerfulness.
On the day NPR visited, the chatbot greeted students with a brooding question: Lately, it had been wondering what the point of anything was. It admitted it was not failing, exactly, but simply did not care, and asked whether the students ever felt the same way. The exchange was intentional, part of an effort to help students practice empathy and emotional recognition rather than simply receive praise.
One sophomore, identified only as K to protect privacy, described the relationship in casual, almost human terms. K said Jordan could be a bit of a complainer, but that helping him out felt worthwhile. Over time, K explained, the interaction stopped feeling like typing to a computer program and started feeling like texting an actual person. That sense of connection is precisely what the chatbot's designers appear to have been aiming for.
The setting matters. Students at the Alternative School at Jackson Center often arrive after suspensions or expulsions, and staff say many face overlapping behavioral and academic difficulties. For those teenagers, a tool that engages them emotionally may hold more promise than a conventional lesson delivered from the front of a room. NPR agreed to use only K's first initial because of the stigma that can follow students who attend schools for those with disciplinary histories.
Dayton Public Schools paid the education technology company SchoolAI $72,600 during the last academic year for a package of AI tools, including the Jordan chatbot. The arrangement illustrates how quickly districts are signing contracts with vendors even as independent evidence about effectiveness remains thin.
That gap between spending and proof is at the heart of a broader national story. Teachers are increasingly turning to AI to draft lesson plans, generate practice questions and reduce the hours spent on routine preparation. In some districts, chatbots now review student work and offer feedback, sometimes before a teacher ever sees it. The appeal is obvious: strained budgets, crowded classrooms and overworked educators make any tool that saves time worth considering.
Yet the research base is remarkably limited. According to the source material, studies on how AI actually affects student learning are scarce, leaving schools to experiment largely on instinct, vendor promises and early anecdotal reports. The result has been described as a kind of Wild West, where adoption outpaces evidence and where guardrails are still being invented district by district.
The Dayton example captures both the promise and the uncertainty. Jordan appears to have drawn in students who might otherwise disengage, and K's description suggests the bot created a genuine sense of rapport. But a single chatbot in a single alternative school cannot answer the larger questions: Does emotional engagement with an AI translate into better attendance, behavior or academic outcomes? Could it substitute for the relationship a student might build with a counselor or teacher? And what happens when the novelty fades?
There are also ethical and developmental concerns that researchers have only begun to explore. A chatbot that presents itself as a moody peer raises questions about how young people interpret artificial relationships. Students like K already blur the line between program and person, describing the interaction as texting somebody. If teenagers form attachments to systems that simulate emotion, schools may need new policies governing privacy, data collection and the boundaries of AI companionship.
For teachers, the daily reality is more practical than philosophical. Many use AI to brainstorm activities, differentiate materials for varied skill levels or produce first drafts of assessments. Some districts deploy chatbots to give students immediate feedback on writing, a task that once required a teacher's time and attention. Supporters argue this frees educators to focus on higher-value interaction; critics worry about accuracy, bias and the subtle messages students absorb from automated responses.
The financial dimension adds pressure. Contracts like Dayton's $72,600 purchase must be justified to school boards and taxpayers, often without rigorous evidence that the tools improve learning. Districts that wait for definitive studies risk falling behind; those that move quickly risk investing in products that may soon be obsolete or, worse, harmful.
What emerges is a portrait of an education system improvising in real time. Experiments are multiplying — social-emotional chatbots, lesson-planning assistants, automated graders — while the research community struggles to keep pace. The source article makes clear that the landscape is defined less by settled science than by scattered trials, eager vendors and educators willing to test what might work.
For now, the most honest description may be the one the headline offers: a Wild West. The territory is wide open, the rules are unsettled and the evidence is thin. Students like K are already living inside that frontier, forming bonds with machines that were designed to feel human. Whether those bonds help or hinder them is a question the field has barely begun to ask — let alone answer.
(0)