As parents and districts push back against AI in the classroom, New Mexico’s struggle to improve student literacy illustrates the high cost of conflating innovation with fear.
In 1991. the United States saw a surge in measles cases that served as a stark reminder of what happens when fear and misinformation override scientific consensus. Decades later. a different kind of public anxiety is taking hold in the nation’s classrooms. and educators fear the consequences for student success may be just as profound.
On July 15. Lynn Vasquez of the New Mexico Public Education Department appeared before state lawmakers alongside Mark Angel. CEO of Amira Learning. Their goal was simple: defend a tool that helps identify reading struggles before they become life-altering deficits. New Mexico has been under a court order since 2018 to improve outcomes for its most vulnerable students. and the state turned to Amira’s AI-powered literacy screener in 2025 as part of that mandate. Unlike generative chatbots that can write essays or hallucinate facts. this software performs a focused. specific task: it listens to 5- and 6-year-olds read aloud. flagging decoding errors and hesitations that teachers might otherwise miss.
Yet, the technology has become a target. In cities like Santa Fe and Albuquerque. parents have organized to opt out of the tool. citing data privacy concerns and a blanket discomfort with artificial intelligence. This local friction gained national momentum in September when New York City—the country’s largest school system—instituted a one-year moratorium on student-facing generative AI for grades 2-K through eighth. Shortly after, Los Angeles pulled similar generative tools from district devices. While these bans target chatbots. the label “AI” has become a catch-all. and many classroom screeners are now being swept into the same defensive bonfire.
This trend creates a dangerous misalignment. The state’s own data shows that the technology is working. When the NMPED shared a report with lawmakers in July. the figures were telling: the Amira screener identified 16.7% of first graders as showing signs of dyslexia. a figure consistent with national estimates. In contrast. the previous computer-based system flagged 45% of students—a massive rate of false positives that often mislabeled children whose first language was not English. By reducing those errors. the current system ensures that finite intervention dollars are directed to the children who actually need help. rather than being diluted by faulty screening.
The human impact of getting this wrong is severe. In New Mexico. where fourth graders currently post the lowest reading scores in the nation. the stakes are not academic; they are developmental. Research commissioned by the Annie E. Casey Foundation indicates that students failing to read on grade level by the end of third grade are four times more likely to drop out of high school. For these students, the “status quo” is not a neutral alternative. It is a pipeline.
Critics of the technology have raised valid questions that demand public answers. Thirty-five New Mexico lawmakers have formally questioned the destination and storage of student voice recordings. and parents in Los Alamos and Santa Fe have expressed fears that recordings could be used to train commercial products. Concerns regarding how AI speech-recognition tools handle various accents or speech differences are also legitimate. as these systems can struggle if not specifically trained for diverse vocal patterns.
However, there is a distinction between governance and obstruction. Albuquerque. for instance. chose a middle path this year: the district retained the screener but established clearer communication with parents regarding the rationale for voice recordings. while simultaneously opting to shorten the time those files are stored.
As the education landscape shifts, the divide between skepticism and panic will define which children receive the support they need. Demanding proof of accuracy, ensuring data privacy, and requiring rigorous vendor testing is responsible governance. But stripping schools of tools that provide an audit trail—and returning to the era of subjective hunches—simply replaces measurable. fixable bias with an invisible one. For the child who cannot read, that choice is a life sentence.
AI in schools literacy New Mexico education reading proficiency student privacy Amira Learning education policy