What happened
After a 2015 hostage-taking and escape attempt at Kaohsiung Prison, Taiwan's Agency of Corrections began building smart surveillance and smart prison systems. They include face recognition and behavior analysis to offset staff shortages. The Chiayi Detention Center received its systems in 2020. In December 2020 the press reported that face recognition took tens of seconds. Members of the Control Yuan, Taiwan's ombudsman body, visited the center on 29 September 2021 and tested the system. In its report approved on 13 April 2022, the Control Yuan found that recognition finished within 3 seconds with a very high recognition rate. It also found that the center's shopping system for inmates cut the time from order to delivery from 7 days to 1 day. The Control Yuan does not say whether that system uses face recognition. At the central control desk, roll calls of a dozen inmates at once worked only when they looked up at the camera. In the same report, the Control Yuan cites an audit finding that staff at seven other prisons confirmed only 29 of 22,601 behavior alerts in September 2019.
Sources
-
Audit report, by authors with no tie to the system.
Control Yuan (監察院), Taiwan. 調查報告(公布版)111司調0010 [Investigation report, public version] on whether the Agency of Corrections' smart surveillance and smart prison projects met their goals. Investigating members 高涌誠 and 王幼玲. Approved 13 April 2022. Investigation opinion 5, p. 14.
另本院於110年9月29日前往嘉義看守所現場履勘,並測試人臉辨識系統,辨識速度確實於3秒之內即可完成,且辨識率極高,該所建置之收容人購物系統,亦對於收容人之生活改善有極大幫助。
Details
AI system
- Workflow
-
The Chiayi Detention Center installed the smart prison systems in 2020 with judicial-technology funds from the Ministry of Science and Technology. The face recognition was one module of a smart surveillance system, beside a module for behavior analysis.
Follow-up
- Conditions for replication
-
Inmates must face the camera when the system checks a group at the central control desk. Camera angle and resolution, which depend on the budget, must support recognition.
Entry details
- Sector
-
Prisons.
- Purpose
-
A vendor that the Control Yuan does not name built the face recognition for the Agency of Corrections' smart prisons, to identify inmates and offset staff shortages.
Not yet checked: model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, AI contribution, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.
Cite this record
Peter H. Vartanian. (2026). Face recognition at a Chiayi jail works within 3 seconds in a test (AIBD-0704). Artificial Intelligence Benefits Database, edition of 31 October 2026. https://petervartanian.xyz/aibd/records/AIBD-0704.html (archived at https://doi.org/10.7910/DVN/AADSII, version 4.0)