School districts are pushing back against state-mandated artificial intelligence tools in classrooms. In New Mexico, superintendent decisions to pause an AI-driven reading app show deep anxieties over how student biometric data, like voice recordings, is collected and stored. These moves reflect a national debate over student privacy and the lack of independent research backing commercial classroom technology.
What Happened
In August 2026, New Mexico’s push to mandate digital reading assessments stalled. Santa Fe Public Schools Superintendent Christine Griffin announced that the district would pause using Amira, an AI reading coach, until state officials provide clearer data privacy guidelines. The Los Alamos Public School District also rejected the state requirement for similar reasons. Albuquerque Public Schools, however, confirmed it will continue using the software to evaluate third graders.
Similar concerns surfaced in Oregon. In June 2026, Portland Public Schools board members raised alarms over a pilot contract with Amira that listed potential collection of sensitive demographic data, including student birthplaces, disability status, and foster care history. Amira's Chief Operating Officer Amy Scholz issued a formal letter clarifying that the company does not receive, store, or require this data to operate. Still, the incident increased worries about the scope of vendor contracts.
The Bigger Picture
Amira Learning is widely used in educational technology. Following its merger with digital learning provider Istation in 2024, the platform is now distributed by Houghton Mifflin Harcourt and serves more than 1,800 school districts globally.
Despite its size, independent evidence of Amira's academic results is scarce. The company promotes marketing materials claiming the platform delivers a 0.40 effect size on reading growth, which it describes as equivalent to high-dosage human tutoring. However, researchers note this is a vendor-reported claim that lacks independent verification. Boston University's EVAL Collaborative selected Amira for its EdTech Evaluation Challenge in March 2026, but the randomized controlled trial evaluating the software is still running.
Teachers have also identified operational problems. The AI voice recognition system struggles to accurately evaluate students with soft voices, lisps, accents, or language barriers. This leads to worries about unfair grading and inaccurate academic labeling.
What This Means for Families
As previously reported, how school AI tools are designed and used matters more than banning them. For parents and educators, the quick adoption of these programs creates two distinct challenges.
First, there is a risk of bias. When an AI tool misinterprets a child's speech pattern or accent, it can flag them as a struggling reader. Teachers then have to manually override the incorrect assessment.
Second, biometric data privacy is unclear. While Amira's corporate documentation details how website cookies manage traffic, there is little public information on how the company handles and deletes audio recordings of children's voices.
What You Can Do
- Review your district's software contract by asking administrators for the explicit data-sharing agreement (DSA) with Amira to see what student information is shared.
- Inquire about parent opt-out options. If you do not want an AI program recording your child's voice, ask for traditional, teacher-led paper assessments instead.
- Request a manual override if your child receives an unexpectedly low score. Ask the teacher to conduct a human-administered reading check to ensure accuracy.