evidoria

← Back to browse

Good practice Imported

Escola Paraná Biometria — ANPD Suspends Facial-Recognition Attendance Tracking for 1 Million Students

Brazil · Curitiba · See the Brazil profile · See the Curitiba profile

Evidence: Observational / pre–post Top 100% 7/100 · Ask Evidence Copilot about this practice

Paraná's education department used facial recognition for attendance in over 2,000 public schools, covering about 1 million students. Brazil's ANPD suspended the practice in August 2026, citing high-risk biometric processing of children without adequate legal basis.

1600 schools
Schools using the system at peak
1 million
Students covered
91.1 %
System accuracy (independent Unesp test)
95 %
Contractual accuracy threshold
Escola Paraná Biometria — ANPD Suspends Facial-Recognition Attendance Tracking for 1 Million Students

Details

Maturity
Discontinued
Promoter
Secretaria de Estado da Educação do Paraná (SEED-PR), Celepar and Valid Soluções
Period
2023 – August 2026
Keywords
K-12 public education, biometric attendance, data protection

Context

Since 2023, Paraná state's education department (SEED-PR) used a facial-recognition system built with state IT company Celepar and vendor Valid Soluções to record classroom attendance via the Escola Paraná app, matching photographed classes to registered student photos; at peak the tool was used in more than 1,600 (later reports ~2,100) of the state network's schools, covering approximately one million students.

Objectives

Automate classroom attendance-taking across the Paraná state school network using facial recognition.

Activities

In April 2025, a Ministério Público do Paraná (MPPR) prosecutor filed a civil action against the state government, Celepar and Valid Soluções seeking suspension and collective moral damages; independent testing by São Paulo State University (Unesp) researchers found the system's accuracy at 91.1%, below the 95% threshold specified in the state's own procurement contract.

Results

On 4 August 2026, Brazil's national data protection authority (ANPD) ordered the immediate suspension of the system, classifying biometric processing of children's data as high-risk and citing five groups of irregularities, including absence of an adequate legal basis under Brazil's LGPD (Articles 11, 14 and 46-47), insufficient security measures, and failure to prioritise the best interests of the children whose data was processed.

Conclusions

The case is recorded as a large-scale, well-documented cautionary example - evidence of harm and regulatory failure at national-programme scale, not of a positive learning or governance practice.

Implementation
Indicative cost
High (€500k–€5M)
Time to results
Long (> 3 years)
Staffing & skills
Secretaria de Estado da Educação do Paraná (SEED-PR), Celepar (state IT company), Valid Soluções (vendor), Teachers operating photo-capture via the Escola Paraná app

Conditions for success

  • A valid legal basis for biometric processing of children's data under LGPD
  • Accuracy meeting or exceeding the contractual 95% threshold
  • Security and data-governance measures adequate for high-risk biometric processing

Common failure modes

  • Absence of an adequate legal basis under LGPD (Articles 11, 14, 46-47), per ANPD findings
  • Measured accuracy (91.1%) below the state's own 95% contractual threshold
  • Insufficient security measures and failure to prioritise children's best interests, per ANPD's suspension order

Where it fits

Governance type
state government + state IT company + private vendor
Scale
state-wide (Paraná), ~1 million students
Income level
upper-middle-income (Brazil)

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