AI Reading Tutor Paused in New Mexico Over Student Privacy Fears

New Mexico school districts are halting a state-mandated AI reading tool, Amira, due to concerns over how student voice recordings are stored and used.

Friday, August 14, 2026

Key Takeaways

  • Two New Mexico school districts, Santa Fe and Los Alamos, recently stopped using an AI reading tutor called Amira. The state requires the software, but local officials paused its use due to student data privacy concerns.
  • Federal privacy laws do not protect student data from AI developers. Frameworks like the Family Educational Rights and Privacy Act (FERPA) and the Children's Online Privacy Protection Act (COPPA) fail to regulate AI model training directly. Without specific state laws, companies can use student voice recordings to train commercial AI models.
  • Some states offer stronger protection. California’s Student Online Personal Information Protection Act (SOPIPA) bans companies from creating student profiles, which limits how developers can use school data.

Two New Mexico school districts have paused the use of a state-required artificial intelligence reading tool over concerns about student data privacy. The decision follows growing anxiety among parents and educators regarding children's voice recordings in classroom technology.

What Happened

Superintendent Christine Griffin of Santa Fe Public Schools announced that the district will not use Amira, an AI-powered reading assistant, including its voice-recording features. As reported by the Albuquerque Journal, both Santa Fe and Los Alamos school districts paused the software's rollout until they receive clearer guidance from the New Mexico Public Education Department. Other districts, including Albuquerque Public Schools, are proceeding with the tool as mandated by the state.

The pause is driven by privacy concerns and practical issues with the software. According to NBC News, some parents report that young children struggle with the program. For example, one first-grader said the assistant regularly interrupted her and did not let her finish sentences, which frustrated her instead of helping her read.

The Bigger Picture

Amira listens to students read aloud and identifies errors to provide automated feedback. While its creators claim the tool offers personalized tutoring that humans cannot easily replicate, critics argue the backend processes are unclear. A review of Amira's privacy policies shows that the company tracks website data using cookies from Google, YouTube, and HubSpot, but offers little clarity on how student voice recordings are stored or if they are used to train proprietary AI models.

This opacity shows the conflict between modern AI capabilities and outdated federal privacy laws. The Family Educational Rights and Privacy Act (FERPA), passed in 1974, was designed for paper records rather than neural networks. Under FERPA’s "school official exception," schools can share data with tech vendors without parental consent, but only if the vendor remains under the school's direct control. When an EdTech vendor retains student records to train its AI models, it violates federal data deletion rules.

Putting student information into AI systems also creates what cybersecurity experts call the prompt exposure problem, where sensitive data is sent to uncontrolled, general-purpose AI servers. As we wrote in our analysis of why the shape of school AI tools matters, school systems must focus on how these platforms are structurally designed rather than relying on outright bans.

What This Means for Families

The debate over Amira shows that student data privacy now includes biometric data like voice prints, rather than just social security numbers or grades. Because federal regulations like FERPA and the Children's Online Privacy Protection Act (COPPA) do not fully regulate how AI vendors use data downstream, families must rely on state laws.

States like California have enacted the Student Online Personal Information Protection Act (SOPIPA), which prevents EdTech providers from building non-educational student profiles or selling student data. Without these state laws, parents cannot easily know if their child’s voice is being used to train commercial AI systems.

What You Can Do

Parents can take several steps to protect their children's data. First, ask your school district's administration for the contract signed with the AI vendor to see if student data is excluded from machine learning training models. Next, check if your state or local school board allows parents to opt children out of voice-recording assessments in favor of teacher-led evaluations. Finally, observe your child when they use the software. If the program's pacing or interruptions cause stress, report these issues directly to school leaders.

Share: