An 'AI school' that never opened and a director on the payroll: the cost of announcing before governing the technology

By Hoodline · November 9, 2026.

Caleb Haraguchi-Combs, 33, is still earning $192,374 a year as the designated principal of a high school that never opened, according to the New York Post, as reported by Hoodline. The school, Next Generation Technology High School, was meant to train students in cybersecurity, computer science, robotics and math in Manhattan's Financial District. Its opening hinged on approval by the Panel for Educational Policy on April 29, and that approval never came: Schools Chancellor Kamar Samuels withdrew the proposal in April.

According to the Post, Haraguchi-Combs received a raise of about $50,000 in January when he took on the role of principal for the project. He previously earned $141,960 as an assistant principal at an East Harlem high school, where he no longer works. Having never actually served as principal, he is not eligible for tenure. Department of Education sources cited by the paper say that in such cases the usual practice is to return the person to an assistant principal post.

The city's Department of Education told the Post that Haraguchi-Combs is working on new school proposals and supporting others that are still underway. According to the paper, he declined to say where he currently works. These are claims from the press and anonymous sources: nothing in the material indicates any wrongdoing on his part, and the mismatch being highlighted is more administrative than personal.

The context matters more than the salary. According to Chalkbeat and Gothamist, the school would have replaced a small all-girls school, and District 2 families criticized the location, the admissions system and the lack of consultation. Gothamist adds that there were concerns about plagiarism, data security and the impact of AI on cognition and mental health. It also noted that the city's selective schools tend to have more white students.

**Our take.** Two stories are tangled together here. The first is about public management: appointing and paying a leader before the project is approved is a recipe for waste, as internal sources cited by the Post acknowledge. The second matters more in the long run: according to critics quoted by Gothamist, part of the backlash against the 'AI school' stemmed from the school system having been slow to set clear rules on AI use.

In general, rolling out technology ahead of its governance tends to be costly. Parents were not so much opposed to robotics as to not knowing what would be done with their children's data or how plagiarism would be prevented. It is a familiar pattern: adoption moves faster on the supply side than in the trust of those on the receiving end.

It would be wrong to conclude that AI has no place in the classroom. In the short term, the transition in education will be clumsy, with failed projects, legitimate resistance and money poorly spent. But in the long term, the augmented tutor and personalized learning may be one of the most powerful levers toward a society of abundance where everyone can devote themselves to what they love. Getting there requires rules on data, assessment and equitable admissions before cutting the ribbon, not after.

The evergreen lesson: an educational AI project must first earn families' trust, with transparency about data, open admissions criteria and public success metrics. Otherwise, it ends up like this case: a name on the payroll and no classroom open.