A Two-Year Investigation Ends in Suspension, Not Prohibition
On August 4, 2026, the Superintendência de Fiscalização of Brazil's National Data Protection Authority (ANPD) issued Despacho Decisório nº 2/2026/SFI, ordering the state of Paraná to immediately suspend all biometric processing of children's and adolescents' facial data in its public school system. The order caps an investigation that began from a citizen complaint and ran nearly two and a half years, culminating in Nota Técnica nº 4/2026/CPDP/CGF/SFI/ANPD, which found the state's Education Department could not produce a valid legal basis under Article 11 of Brazil's LGPD (Lei Geral de Proteção de Dados) for processing sensitive biometric data of minors, and that the system's design failed the best-interest-of-the-child standard under Article 14.
The scale is what makes this a landmark case rather than a routine compliance action. "Escola Paraná Biometria," piloted in 2021 and rolled out statewide in 2023, ran facial recognition across 2,136 schools, roughly one million students, and more than 100,000 staff, making it likely the largest deployment of biometric surveillance on minors documented anywhere. Teachers photographed students each class period through an app built by the state's own tech company, Celepar, with images processed by contractor Valid Soluções using facial-matching software supplied by the Slovak firm Innovatrics. Attendance data fed directly into eligibility tracking for Bolsa Família, Brazil's flagship conditional cash-transfer program, which covers more than 20 million families and disbursed R$12.8 billion in April 2026 alone.
The Case for Caution Here Is Real
Before dismissing this as regulatory overreach, it's worth taking the strongest version of the case for the system seriously. Manual attendance-taking in a system this large is genuinely error-prone and administratively costly, and accurate attendance is the actual gatekeeping mechanism for a welfare program that lifts millions of children out of poverty — a legitimate public interest, not a pretext. Automating it isn't inherently sinister; plenty of jurisdictions use biometric tools for narrower, better-governed purposes (airport boarding, banking KYC) without controversy. If Paraná had built this with a documented legal basis, a real data protection impact assessment, parental consent mechanisms, and independently verified accuracy, the debate would be about proportionality at the margins, not about baseline legality.
That's precisely why the ANPD's finding matters: none of those things existed. The agency's technical note cataloged five failures — no valid legal hypothesis under Article 11, a disproportionate and unnecessary processing method for what a sign-in sheet or ID scan could accomplish, inadequate security and access governance, no assessment of less invasive alternatives, and non-cooperation with the investigation itself. An independent study by UNESP found the system's accuracy sat at 91.1%, below the 95% contracted threshold, on a program whose stated purpose was tying attendance to welfare eligibility for low-income families. Comparable systems were already blocked elsewhere in the world on similar grounds: Sweden's DPA found a nearly identical school attendance-scanning system violated the GDPR in 2019, and a French administrative court in Marseille ruled in 2020 that children cannot meaningfully consent to biometric monitoring given the power imbalance inherent in a classroom.
Suspension, Not Ban — And That Distinction Should Hold
The ANPD was careful to frame this as a preventive measure, not a sanction — the despacho is explicit that it precedes, and could still avoid, formal administrative penalties, and it gives Paraná ten business days to certify compliance. Paraná's Education Department, for its part, told reporters the pilot had already been discontinued in July 2026 with the Valid Soluções contract terminated — before the ANPD order landed, though after the two-year investigation had made suspension all but inevitable. A parallel civil suit filed by the Paraná Public Ministry (Ação Civil Pública nº 0004208-55.2025.8.16.0058) remains active in state court, with Data Privacy Brasil and Privacy International filing as amicus curiae.
This is the right outcome, and the right instrument. The ANPD did not categorically prohibit biometric attendance technology in Brazilian schools; it enforced the existing legal basis and necessity requirements the LGPD already imposes, and found this specific implementation failed on documented, technical grounds — not because facial recognition touched children, full stop. That distinction matters for innovation policy: a categorical ban on biometric identification in education would foreclose future systems that might clear a proportionality bar Paraná's never approached — narrower scope, verified accuracy, real consent, and a legally defensible purpose. Pending legislation, Bill No. 1.225/2026, would go further by statutorily barring any automatic link between biometric attendance and welfare eligibility — a more durable fix than case-by-case enforcement, since it addresses the specific harm (an unreviewable false negative cutting off a family's benefit) rather than the technology category. Brazil's regulators enforcing the law as written, rather than reaching for a blanket prohibition, is the model other jurisdictions weighing biometric school surveillance should be watching.