evidoria

← Back to browse

Good practice Imported

DRC Cuts National State-Exam Correction from 45 Days to 3 with AI-Assisted Scanning

Congo (the Democratic Republic of the) · Kinshasa · See the Congo (the Democratic Republic of the) profile · See the Kinshasa profile

Evidence: Observational / pre–post Top 80% 33/100 · Ask Evidence Copilot about this practice

The DRC's Ministry of National Education introduced AI-assisted scanning and correction for the 2025 State Exam, adding scanning centres in Mbuji-Mayi and Lubumbashi and cutting the correction cycle from 45 days to 3, with human inspectors retaining oversight.

45 days
Exam correction time before AI-assisted scanning
3 days
Exam correction time after AI-assisted scanning (2025 EXETAT session)
92750 candidates
Candidates in Bandundu-ville (2025)
44355 candidates
Candidates in Bukavu (2025)
50.2 %
Share of girls among Bukavu candidates (2025)
3000 centres (more than)
Examination centres nationwide (2025)
35 % (below)
National internet penetration (2024)
5 of 26 provinces
Provinces with concentrated internet access (2024)
DRC Cuts National State-Exam Correction from 45 Days to 3 with AI-Assisted Scanning

Details

Maturity
Established
Promoter
Ministère de l'Éducation nationale et Nouvelle Citoyenneté (RDC) — Examen d'État (EXETAT)
Period
2025 State Exam (EXETAT) session, ongoing
Keywords
K-12 education, national examinations, public administration, AI-assisted grading

Context

The Democratic Republic of the Congo's Ministry of National Education and New Citizenship modernised correction of the 2025 Examen d'État (EXETAT), the national secondary-school leaving exam, introducing AI-assisted scanning and automated correction alongside continued human review.

Objectives

To speed up and secure the marking of a national exam sat by hundreds of thousands of candidates, while curbing human error and fraud, and retaining human oversight of the automated process.

Activities

Minister of State Raïssa Malu launched a modernised correction centre at the Don Bosco school complex in Kinshasa's Masina commune; two further modernised scanning centres opened in Mbuji-Mayi and Lubumbashi for the 2025 session, digitising scripts locally before sending them to Kinshasa for correction under the supervision of General Inspectorate experts. An early correction start from the second day of the exam in several centres was cited as a key factor behind faster results.

Results

Correction time fell from 45 days to 3 days for the 2025 EXETAT session, which ran across more than 3,000 examination centres nationally; regional figures cited include 92,750 candidates in Bandundu-ville and 44,355 in Bukavu (50.2% girls). The Ministry is also introducing blockchain-secured e-diplomas as part of a wider digital push.

Conclusions

The initiative shows a measurable administrative efficiency gain at national scale, but the Ministry itself acknowledges major unresolved equity gaps — national internet penetration was below 35% in 2024, concentrated in only 5 of 26 provinces, and reaching remote and conflict-affected areas remains unaddressed.

Implementation

Indicative cost
Medium (€50k–€500k)
Time to results
Medium (1–3 years)
Staffing & skills
Ministry of National Education and New Citizenship officials, General Inspectorate experts (correction oversight), Local scanning-centre technical staff in Kinshasa, Mbuji-Mayi and Lubumbashi

Conditions for success

  • Retaining human oversight of AI-assisted correction via General Inspectorate experts
  • Local digitisation/scanning centres to reduce transport time before central correction
  • Early correction start (from exam day two) in several centres

Common failure modes

  • Internet penetration below 35% in 2024, concentrated in only 5 of 26 provinces, limiting reach to remote and conflict-affected areas
  • No published data-protection policy, error-audit methodology or appeals process for contested results

Where it fits

Governance type
national government ministry
Scale
national

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