Sheng Siong's facial recognition catches a repeat wine thief

Sheng Siong's AI-driven facial recognition closed-circuit television system: a suspect's photograph, taken from CCTV footage after a theft is found, is uploaded, the system matches the face across footage and sends mobile alerts through a corporate app to staff when that person enters any Sheng Siong outlet

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

A stock check that revealed missing wine led staff to review camera footage and upload the suspect’s face. The system sent an alert that reached the manager when she returned the next day. Staff stopped her at the checkout. Police and court proceedings that followed led to restitution and an eight-day sentence.

Staff found the initial discrepancy, reviewed the footage, and uploaded the photograph. A manager intercepted the customer and contacted the police.

The system's documented contribution came after a human detection. On 10 September 2025 a staff member found discrepancies in a stock check. Colleagues reviewed CCTV and saw Tan leaving without paying. They uploaded her face. The Straits Times reports that the system “then automatically identified the other instances” of theft. That turned one observed theft into an amalgamated charge covering seven visits. On 11 September the system's alert reached the retail manager's phone when she entered. He intercepted her at the self-checkout kiosk with three unscanned bottles. The interception, the bag check, the police call, the prosecution and the restitution were human and institutional. No source names the vendor or the model, gives an accuracy figure, or reports a false match in this case.

After staff identified a theft, facial-recognition software found footage of earlier visits and alerted a manager when the same person returned to the store.

Not assessed

Sheng Siong recovered the value of stolen wine and stopped three further bottles from leaving the store.

Reduced Violence

Not assessed

Organizations.

Sheng Siong’s Punggol Central store.

Not assessed

Not assessed

Not assessed

Yes

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

Bottles recovered, restitution paid, and the interval between identifying the suspect and the next alert.

Assumed alternative: the chain's previous practice of staff manually reviewing CCTV footage to identify suspects

The alert reached the manager when the suspect returned the next day. Staff stopped three unscanned bottles, and the retailer later received about S$470 in restitution. No controlled comparison with manual surveillance is reported.

Not assessed

The sources do not identify the vendor or model and provide no accuracy measure. They do not report a false match in this instance.

Not assessed

Not assessed

restitution paid by the shoplifter for the stolen bottles that were not recovered (three bottles were recovered at the interception)

about S$470

Not assessed

Not assessed

Not assessed

Sheng Siong

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

Operator: Sheng Siong (supermarket chain, 83 outlets), Singapore.

Benefit: Punggol, Singapore. Operator: Singapore.

Not assessed

Not assessed

Not assessed

Not assessed

Staff must first identify a suspected theft and supply the relevant image. A manager must then verify the match and act before the person leaves. The identification and stop need to be lawful.

1 follow-up entries. Outcome unresolved: 1.

12 May 2026. The shoplifter pleaded guilty to an amalgamated theft charge and was jailed for eight days on 12 May 2026 after making full restitution of about S$470.

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

AIBD-0207-source-0001–0004

Numbered references below.

Not assessed

Independent

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

Not assessed

Organizations.

Organizations: 1.

The store stopped three bottles from being taken and received about S$470 in restitution for earlier thefts. The full series involved 19 bottles worth about S$556. Staff first noticed the missing stock and reviewed the footage, so the documented AI contribution was recognizing the suspect’s return.

Not assessed

Not assessed

Not assessed

Not assessed

Independently reported

Not assessed

Not assessed

Retail loss prevention and law enforcement. Supermarket shop-theft detection: facial recognition on store CCTV with mobile alerts to managers.

Sheng Siong rolled out its AI-driven facial recognition CCTV system in April 2024 as a measure against shoplifters. It worked with the Singapore Police Force under its Shop Theft for Retailers (STAR) program. In August 2025 it announced an extension to all 83 outlets. Staff upload the photograph of a suspect identified from footage. The system alerts store managers' phones when that person enters any branch. Sheng Siong said thefts per store rose from 4.91 in 2023 to 8 in 2024. Its CEO attributed the rise to catching cases the human eye had missed. Identifying a repeat shoplifter and flagging her return was exactly its designed purpose.

Date basis unassessed: 2025 (precision unassessed). Updated: 25 October 2026.

Not assessed

SHA-256: 692dc2995f18…

oecd_aim: 2026-05-12-21c2.

Not assessed

Not assessed

6 revisions. Latest: 25 October 2026.