A Middle School in Chicago Is Using AI to Personalize Every Student’s Learning Path. Early Results Are Remarkable

On a Tuesday morning, most middle schools have the same layout: rows of desks, a teacher at the front, and a whiteboard covered in notes that half of the students are copying and the other half have already tuned out. That is not how Alpha Chicago appears. No rows are present. No instructor is presiding over the court.

An AI system that is silently operating in the background already knows whether a specific twelve-year-old has trouble with fractions, reads two grade levels ahead, and becomes distracted after twenty minutes of screen time. Without anyone having to ask, it makes the necessary adjustments.

Category	Details
School Name	Alpha Chicago
School Type	Private K–12, part of Alpha Schools network
Founded	2014
Founder	Mackenzie Price
Location	Chicago, Illinois, USA
Model	AI-assisted personalized learning with human “guides”
Daily AI Use	1–2 hours per day for core subjects
Subjects Covered	Math, Science, Reading
Annual Tuition	$55,000
Student Performance	Top 1% on national standardized tests
Learning Rate	2.6x faster than peers on MAP assessments
Current Enrollment	2 enrolled; 35 expressed interest
Target Enrollment	50 students by fall 2026
Total Alpha Schools (US)	22 campuses nationwide
Teacher Role	Guides, not replaced — paid high salaries
Academic Framework	Montessori-influenced, self-directed learning

The school is a member of Alpha Schools, a private K–12 network with 22 campuses across the US that Mackenzie Price founded in 2014. This autumn, the Chicago location is getting ready to dive deeper into AI, utilizing it for one to two hours every day in core subjects like science, math, and reading while human “guides” move around the classroom, observing, prodding, and stepping in when something doesn’t work.

Price has resisted the easy headline with caution. “We are using the same curriculum that students in the classroom are learning from,” she said to CBS. “This is not ChatGPT coming up with made-up questions.” That distinction is more important than it may seem.

An automated system that spits out content without accountability and confuses engagement metrics for real learning is one type of AI in education that, understandably, causes anxiety. Price’s description sounds different: a diagnostic engine that determines where a student’s comprehension falters and then reacts in real time by modifying the pacing, difficulty, and format.

That’s not a novel concept. For decades, educators have attempted to accomplish this manually, with varying degrees of success, due to the fact that a single teacher can only monitor thirty children at once.

Even though the figures cited by Alpha Chicago merit careful consideration, they are difficult to discount. On MAP growth assessments, students advance at a rate about 2.6 times faster than their peers and place in the top 1% of the country on standardized tests. Liz Gerber of the Center for Human-Computer Interaction and Design at Northwestern University put it this way: “It’s really not that new, to be honest, it’s personalized learning.”

She is correct that the idea is outdated. The technology that powers it is now different: AI that can create and update a learner profile, identify gaps before they become failures, and recommend the best challenge for a particular student at a given time. A century ago, the Montessori movement used wooden blocks and colored beads to achieve similar objectives. The tools have undergone significant change.

However, it’s difficult to ignore who this model is currently working for. Alpha Chicago attracts a select group of Chicago families who can afford to pay the $55,000 annual tuition. Although 35 students have expressed interest and the school hopes to have 50 by fall 2026, only two have actually enrolled for the upcoming academic year.

These are tiny numbers, and the model hasn’t yet been put to the test on a large scale or with groups that have historically benefited the least from traditional education. Any sincere enthusiasm must stop there. The most pressing question is not whether wealthy families can access personalized AI learning—they already can—but rather whether it will ever reach the children who need it the most if it performs as promised.

Price maintains that the teacher’s question was misinterpreted from the beginning. “Teachers are not going to be replaced,” she declared. “They are the most important part of making a model work.” They are referred to as guides at her school and are well compensated for their work, which is more about observing the people in the room and identifying who is irritated, who has checked out, and who requires a different kind of support than an algorithm can offer. It’s intriguing that a school built around automation is so overtly reliant on it. That’s a real skill, perhaps even more difficult to master than traditional instruction.

In other places, models like Squirrel AI in China, Carnegie Learning, and Summit Public Schools in California have been subtly advancing the idea that adaptive learning systems can make a difference, particularly when teachers are truly included in the process rather than excluded from it.

The evidence is not perfect, and there are still unanswered questions regarding algorithmic bias, data privacy, and what is lost when education becomes data-driven. Not every child learns in a way that can be measured by a system.

Nevertheless, something seems to be changing. Not everyone found success in the previous classroom; it was simply failing more subtly. Located in a single area of a city, Alpha Chicago is a small, costly, early-stage experiment. However, it involves asking the right questions, which is less common than it ought to be in the field of education. What the coming years will show is whether the answers scale.