What happened
Storstockholms Lokaltrafik (SL), Stockholm's regional public transport authority, installed AI-CCTV at 14 of its 100 metro stations in October 2021. The software, IRIS-SL by Irisity, scans the feeds of existing platform cameras. It raises an alarm when a person enters the track area or stands at the platform edge. A safety officer reviews each alarm, and controllers can warn drivers, cut the power or call emergency services. Researchers at Karolinska Institutet compared these stations with the other 86 in a controlled interrupted time series. In the following year, people hit by trains in suicidal acts fell from 1.17 to 0.50 a quarter at the 14 stations. At the other stations the rate rose from 2.30 to 3.00. The primary controlled analysis gave a rate ratio of 0.32. No suicides occurred at the 14 stations that year, against six at the others, too few for statistical significance. Canceled train-kilometers also fell. By mid-2025, SL had logged 36 alarms that led to a person being safeguarded. The Swedish Transport Administration funded the study. Three of its authors also work for Stockholm Health Care Services, a part of Region Stockholm, which also runs SL. The authors call the findings preliminary because the follow-up lasted four quarters and events were few.
Sources
-
Peer-reviewed study, by authors with no tie to the system.
Fredin-Knutzén J, Hadlaczky G, Andersson A-L, Sokolowski M. Reduction in suicides and suicide attempts following implementation of AI-based video surveillance in the Stockholm metro system: an intervention study. BMC Public Health. 2026;26:1997. doi:10.1186/s12889-026-28358-1. Published 1 July 2026. Abstract, Results and Discussion. Additional file 1, section AI-CCTV, names the software.
Implementation of AI-CCTV as described herein was associated with lower rates of PUT due to suicidality at metro stations, as well as changes in the same direction for e.g. deaths by suicide and less cancelled train-kilometers
-
Press release, by the makers.
Irisity AB. Irisity vinner SLs omfattande upphandling av videoanalys för tunnelbana (Irisity wins SL's large procurement of video analytics for the metro). Press release via Cision, 27 November 2020.
Irisity AB (publ) har valts av AB Storstockholms Lokaltrafik att leverera deep learning algoritmen IRIS Rail till Stockholms tunnelbana.
Details
AI system
- Workflow
-
The software scanned the feeds of existing platform cameras and raised an alarm when a person entered the track area or stood at the platform edge. A safety officer reviewed each alarm and passed elevated risks to a train traffic controller. The controller warned the train driver, cut the power and called emergency services.
Follow-up
- Conditions for replication
-
Staff can slow or stop a train before it reaches the person. Cameras cover the platform near the track at enough resolution to detect a person.
Entry details
- Sector
-
Public transport safety / suicide prevention.
- Purpose
-
Irisity built IRIS-SL to scan station camera feeds and alert SL's safety officers when a person enters the track area or stands at the platform edge.
30 more questions not yet checked
model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, 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). Fewer suicidal acts on the tracks at Stockholm stations with AI-CCTV (AIBD-0953). Artificial Intelligence Benefits Database, edition of 1 November 2026. https://petervartanian.xyz/aibd/records/AIBD-0953.html (archived at https://doi.org/10.7910/DVN/AADSII, version 7.0)