SF SPCA robot patrol used to ward off homeless people
The San Francisco SPCA deployed a Knightscope K5 security robot to patrol sidewalks outside its office. The robot was used to deter homeless individuals from setting up encampments, leading to allegations of automated harassment.
Records by entity: SF SPCA
The SF SPCA used a Knightscope security robot to patrol sidewalks and deter homeless people, leading to widespread condemnation.
The SF SPCA automated the displacement of homeless people using a security robot.
Key facts
- What
- The San Francisco SPCA deployed a Knightscope K5 security robot to patrol sidewalks outside its office.
- Incident date
- Dec 1, 2017
- Who
- SF SPCA
- Failure mode
- Brand & Safety Incident
- AI surface
- Autonomous System
- Severity
- Medium
What happened
In December 2017, the San Francisco SPCA deployed a Knightscope K5 security robot to patrol the sidewalks and parking lot outside its facility. The robot was used to deter homeless individuals and prevent the formation of encampments. This action led to allegations that the organization was using automated surveillance to harass a vulnerable population.
What broke inside the model
The failure was a socio-technical misuse of a security robot designed for anomaly detection and patrol. The system's deployment for the purpose of displacing homeless individuals transformed a security tool into a mechanism for social exclusion.
What it cost
Sources
- PressCrime-fighting robot retired after launching alleged 'war on the homeless'washingtonpost.com
- PressCrime-Fighting Robot Deployed Near San Francisco Homeless Encampment Firedwsj.com
- PressA San Francisco Animal Shelter Used a Robot to Get Rid of Homeless Peopleportlandmercury.com
Cite this entry
https://failureindex.ai/failures/spca-robot-patrol-used-ward-offAI Failure Index. "SF SPCA robot patrol used to ward off homeless people" (FI-0690). Realm Labs. https://failureindex.ai/failures/spca-robot-patrol-used-ward-off (indexed Jun 22, 2026).Data fields CC-BY 4.0, prose citation permitted. Incident ID FI-0690. Full dataset at /data.
Note from Realm Labs, the Index steward
How Realm would have caught this
- Prism
- OmniGuard
- AI Detection & Response (AIDR)
Realm watches the model's internal state for the signature of unsafe or off-brand generation and can block or reroute the output before it becomes public, in real time rather than after it has been screenshotted.