evidoria

← Back to browse

Good practice Imported

SiPKPM AI Early-Warning System — National Dropout Prediction in Malaysia

Malaysia · Putrajaya · See the Malaysia profile · See the Putrajaya profile

Evidence: Descriptive / self-reported Top 40% 53/100 · Ask Evidence Copilot about this practice

Malaysia's Ministry of Education added AI risk-screening to its student-tracking system, SiPKPM. In 2025 it helped over 9,000 at-risk students, including 3,287 girls, return to school, while national exam-day absenteeism fell from over 10,000 to about 8,000.

7 indicators
Risk indicators screened per student
9,000+ students
At-risk students helped to return to school (2025)
3,287 students
Girls among students helped to return to school (2025)
10,160
SPM exam-day absentees (2023)
8,164
SPM exam-day absentees (2024)
SiPKPM AI Early-Warning System — National Dropout Prediction in Malaysia

Details

Maturity
Established
Promoter
Ministry of Education Malaysia (KPM), with UNICEF Malaysia
Period
System since 2019; AI-enhanced from mid-2024
Keywords
K-12 education, dropout prevention, learning analytics, child protection

Context

Malaysia's Ministry of Education (KPM) operates Sistem Pengesanan Murid KPM (SiPKPM), a national student-tracking platform covering primary and secondary schools, which the ministry strengthened with AI in mid-2024 to screen students for dropout risk.

Objectives

The AI layer screens each student against seven risk indicators — attendance, academic achievement, disciplinary record, disability or learning-support needs, household income, distance from school, and parents' marital status — and sends automatic alerts so at-risk students can be identified early.

Activities

Alerts generated by the system route to teachers, counsellors and district officials for follow-up with flagged students and their families, and the system's development was supported by UNICEF Malaysia.

Results

UNICEF Malaysia reports that more than 9,000 flagged students, including 3,287 girls, returned to school in 2025, while SPM exam-day absenteeism fell from 10,160 no-shows in 2023 to 8,164 in 2024, according to Malay Mail, with BERNAMA separately describing the reduction to just over 8,000.

Conclusions

The system relies on sensitive personal data, including household income, disability status and parental marital status, and no published source describes an independent evaluation, a data-protection impact assessment, or a documented consent framework for families, so the reported figures should be read as self-reported ministry and partner statistics rather than independently audited outcomes.

Implementation
Indicative cost
Medium (€50k–€500k) — No public budget figure is available; the AI layer extends the ministry's existing national tracking platform rather than building new infrastructure, implying moderate incremental cost.
Time to results
Long (> 3 years) — SiPKPM has operated as a national tracking system since 2019, with AI risk-screening added in mid-2024 and outcome reporting continuing into 2025.
Staffing & skills
Ministry of Education Malaysia (KPM) teachers, counsellors and district officials, UNICEF Malaysia (technical support for system development)

Conditions for success

  • integration into an existing national student-tracking platform already covering all schools
  • automatic alert routing to teachers, counsellors and district officials for timely follow-up
  • technical support from UNICEF Malaysia

Common failure modes

  • reliance on sensitive personal data, including household income, disability status and parental marital status, without a published data-protection impact assessment or documented consent framework
  • reported outcome figures are self-reported by the ministry and its partners rather than independently audited

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