ScamShield flags about 5 million SMSes by October 2023

AI system

System: ScamShield app
Model: Not assessed
Release: Not assessed
Application developer: Open Government Products (GovTech) with the National Crime Prevention Council and the Singapore Police Force
Role: Not assessed
Configuration: Not assessed
Inputs and outputs: Not assessed
Autonomy: Not assessed
Workflow: On-device SMS classification combined with a police-maintained blocklist of scam numbers. User reports feeding the blocklist (722,865 SMS reported by mid-2021).
Human oversight: Not assessed
Relation to AI: GovInsider attributes SMS filtering to an on-device machine-learning model and call blocking to a police-maintained blocklist. So only the SMS figures involve the model. The parliamentary reply gives counts without attributing them to AI. Developer (Open Government Products, a public body) and deployer (the Police and NCPC) are both government bodies. The reply was given by their ministry.
AI contribution: GovInsider attributes SMS filtering to an on-device machine-learning model and call blocking to a police-maintained blocklist. So only the SMS figures involve the model. The parliamentary reply gives counts without attributing them to AI. Developer (Open Government Products, a public body) and deployer (the Police and NCPC) are both government bodies. The reply was given by their ministry.

Studies: Not assessed
Study design: Not assessed
Measurement: Not assessed
Setting and sample: Not assessed
Measures: about 5 million SMSes had been flagged to users as potential scams
Comparison: Not assessed
Results: about 5 million SMSes had been flagged to users as potential scams
Additional value from AI: Not assessed
Limitations: Not assessed
Denominators: Not assessed
Materials for replication: Not assessed

Participants: Developer: Open Government Products (GovTech) with the National Crime Prevention Council and the Singapore Police Force. Operator: Singapore Police Force and National Crime Prevention Council.
Jurisdictions: Benefit: Singapore, Singapore. Developer: Singapore. Operator: Singapore.
Ownership and control: Not assessed
Dependencies: Not assessed
Cross-border connections: Not assessed
Geopolitical implications: Not assessed

Conditions for replication: Adoption at scale (about 750,000 downloads by October 2023). Police intelligence to keep the blocklist current. Model updates as scam wording changes.
Replication attempts: 2 follow-up entries. Benefit recurred: 1. Outcome unresolved: 1.
Later findings: 18 August 2026. MDDI's anti-scam page, updated 18 August 2026, says the enhanced ScamShield app launched in August 2024 grew its user base from 944,000 to 1.53 million.
Availability: Not assessed
Institutional decision: Not assessed
Grounds: Not assessed
Action and result: Not assessed
Success criterion: Not assessed

Assessed benefits: People. Society.
Depth: Not assessed
Reasons: People: Flagged-SMS and blocked-number counts are aggregates. No loss avoided by an identifiable user is documented. Society: No measured change in scam losses is attributed to the app in the sources.
Observed or intended: Not assessed
Rubric: Not assessed
Assessor and date: Not assessed
Alternative assessments: Not assessed
Fill: Independently reported

Kind: Not assessed
Related entries: Not assessed
Field: Scam prevention. Consumer scam calls and SMS.
Purpose: Filter scam SMS on-device and block calls from police-blacklisted numbers to protect users from scams.
Dates: Date basis unassessed: 2023 (precision unassessed). Updated: 21 October 2026.
Stage: Not assessed
Version: SHA-256: e18a9e6a97e4…