Aligned by Design
On May 27, 2026, the American Federation of Teachers called for sweeping limits on screens and AI in schools. We read that guidance as a call we share β and as a description of exactly what CLARO was built not to be.
"CLARO does not ask schools to choose between caution and innovation. It is built to honor both."
What the AFT actually said
The guidance is frequently summarized as "the teachers' union is against AI in schools." That is not what was said. The AFT is asking for three things CLARO can affirm without hesitation: keep teachers central, protect the development of critical thinking, and hold AI tools to real privacy and safety standards. The disagreement in the market is about how, not whether, to protect students.
Source: AFT President Randi Weingarten, "Devices Down, Eyes Up, Hands On" address, National Press Club, May 27, 2026, as reported by The New York Times, Education Week, NBC News, and EdSource.
Where CLARO sits relative to the guidance
Every specific concern the AFT raised maps directly to a design decision CLARO made before the guidance was issued.
| AFT Concern | CLARO's Design Response |
|---|---|
| Screens displacing hands-on learning in early grades | CLARO is a teacher-administered assessment, not a child-facing always-on app. Deployment is configurable by grade band so the youngest learners stay screen-light. |
| Student-facing AI replacing instruction in elementary school | CLARO does not instruct in place of the teacher. Through dialogue with Chispa it surfaces how a student reasons so the teacher can teach more precisely β it informs instruction rather than replacing it. |
| Companion chatbots simulating relationships | Chispa is a dialogue partner for thinking, not a companion. The interaction is a structured cognitive assessment, not simulated friendship or emotional relationship. |
| Chatbots giving "effortless answers" that erode critical thinking | This is CLARO's core inversion. The assessment happens live in dialogue, which AI-proofs student thinking: a student cannot outsource the work to a chatbot. CLARO surfaces "ghost learning" β students who can recall facts but cannot reason with them β measuring and protecting critical thinking by design. |
| Lack of privacy and safety standards | Privacy-first architecture and transparent safety controls are a launch feature, not a roadmap item, consistent with the AFT's call for national standards. |
| Teachers left untrained and sidelined | CLARO puts the teacher at the center as the reader and actor on student reasoning, and builds educator enablement into the adoption model. |
| Student wellbeing and pressure from scores | Students never see DQ levels, scores, or percentiles β only the six Chispa states. The design deliberately protects metacognitive integrity and avoids fixed-mindset reinforcement. |
What we commit to
We would rather be measured against these commitments than asked to be trusted on them.
Teachers stay central.
CLARO is an assessment the teacher administers and acts on. It surfaces how students reason; it does not replace instruction or the relationship between a teacher and a student.
Chispa is a dialogue partner, not a companion.
Students think out loud with Chispa as part of a structured cognitive assessment. It does not simulate friendship or emotional relationships and is not the unsupervised companion use the AFT warns against.
We protect critical thinking.
Because the assessment happens live in dialogue, it AI-proofs student thinking and reading β the work cannot be outsourced to a chatbot. CLARO measures how students reason rather than rewarding effortless answers.
Privacy and safety first.
Strong data protection and age-appropriate controls are part of the product at launch, and we support the call for clear national standards.
Age-appropriate and equitable by design.
We support keeping the earliest grades hands-on and screen-light, configure deployments accordingly, and offer bilingual access so more students get a fair read on their reasoning.
We protect young learners from score pressure.
Students never see levels, scores, or percentiles β only the shared Chispa-state vocabulary β by design, to protect metacognitive integrity and avoid fixed-mindset reinforcement.
The invitation
We invite educators, union leaders, and district administrators to evaluate CLARO against the very concerns the AFT raised β and to hold us to them.
"The future of learning isn't a choice between human teachers and intelligent tools. It's teachers who can finally see how their students think β made more capable by tools built to serve them. That's what CLARO is for."