evidoria

← Back to browse

Good practice Imported

New Mexico's Amira Opt-Out — A Statewide Biometric-Privacy Backlash Against a Mandatory AI Reading Test

United States of America · Santa Fe · See the United States of America profile · See the Santa Fe profile

Evidence: Descriptive / self-reported Top 81% 33/100 · Ask Evidence Copilot about this practice

New Mexico mandates Amira's AI voice-based reading test for all K-2 students, but after parents raised biometric-privacy concerns, six districts including Santa Fe and Los Alamos opted out and the state introduced tiered voice-data-deletion rules.

2.7 USD million
Annual state cost of the Amira mandate
1.4 USD million
Annual cost of the assessment component
6 districts
Districts opting out of the voice-recording feature
3 times/year
Required assessments per year
New Mexico's Amira Opt-Out — A Statewide Biometric-Privacy Backlash Against a Mandatory AI Reading Test

Details

Maturity
Established
Promoter
New Mexico Public Education Department / Amira Learning
Period
2025–2026
Keywords
K-12 education, reading assessment, biometric data privacy, ed-tech procurement

Context

New Mexico's Public Education Department (PED) requires all kindergarten-through-second-grade students statewide to take Amira Learning's AI reading assessment three times a year, plus 30 or more minutes of weekly AI-guided tutoring, at an annual state cost of $2.7 million, of which $1.4 million funds the assessment component alone; Amira listens to children read aloud and analyses the recording to flag errors.

Objectives

Leaders in Santa Fe, Los Alamos, Farmington, Roswell, Clayton and Turquoise Trail Charter School argued that a child's voice is biometric data and objected to a third-party vendor retaining recordings that could be used to train commercial AI models, as well as to documented accuracy gaps for students with speech disabilities or accents.

Activities

Days before the 2025-26 school year, PED responded with an opt-out policy: districts can use Amira without the voice-recording feature, revert to paper-based testing, or choose an alternative assessment, and can request deletion of voice recordings daily, weekly, monthly or at year-end instead of the prior default schedule.

Results

Six districts, including Santa Fe and Los Alamos, opted out of the voice-recording feature over these biometric-privacy and accuracy concerns.

Conclusions

A May 2026 Legislative Education Study Committee brief found New Mexico still lacks statewide AI-procurement standards, AI-specific student data-privacy rules, or a formal evaluation process for education AI tools, meaning the opt-out resolved this one dispute without closing the underlying governance gap.

Implementation

Indicative cost
High (€500k–€5M) — $2.7 million annual state cost, of which $1.4 million funds the assessment component alone.
Time to results
Medium (1–3 years) — Mandate in place through 2025-26; opt-out policy introduced days before the 2025-26 school year; governance gap still open as of the May 2026 legislative brief.
Staffing & skills
New Mexico Public Education Department (PED), District leaders in Santa Fe, Los Alamos, Farmington, Roswell, Clayton and Turquoise Trail Charter School

Conditions for success

  • A concrete, tiered opt-out and voice-data-deletion policy issued in direct response to district objections
  • Districts retaining a choice between the voice-recording feature, paper-based testing, or an alternative assessment

Common failure modes

  • Documented accuracy gaps in the AI reading assessment for students with speech disabilities or accents, raised but not yet resolved
  • As of May 2026, New Mexico still has no statewide AI-procurement standards, AI-specific student data-privacy rules, or formal AI-tool evaluation process

Where it fits

Governance type
state government mandate with district-level opt-out
Scale
statewide K-2 mandate, 6 districts opted out
Income level
high-income

Commonly funded by

National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful