CHAPTER 8
The same streetlight units can serve many uses. But the functions required, and the difficulty of social acceptance, differ greatly by use. We therefore propose adding uses step by step.
| Use | How it works | Additional burden | Social acceptability |
|---|---|---|---|
| Watching over children (Bluetooth) | Streetlights detect a smartphone or small tag carried by a child and record its location | Very small | High; based on consent of the child and parents |
| After-the-fact tracking (basic function) | After an incident or disappearance, people follow images camera by camera | As estimated on this site | To be confirmed in pilots |
| Anomaly detection | On-device AI continuously detects dangerous situations and sends alerts | More power and computation | The most difficult |
| Behaviour-pattern analysis | Trends analysed from accumulated information | Large | Requires careful debate |
Bluetooth-based watching-over rests on consent, so it meets little opposition and costs almost nothing extra. It also improves the accuracy of camera tracking, because "the child carrying this tag" can be linked to the images even in a crowd.
Anomaly detection, with AI judging continuously, is a difficult use to accept. On the other hand, it is "AI output" in the purest sense: what the AI detected, and whom it notified, can all be recorded and audited by VRAIO. We envisage adding it as the final stage, after society has decided.
Units on top of streetlights are not easily replaced. We therefore recommend installing AI-processing parts and Bluetooth receivers from the start, and enabling functions step by step through software updates as society agrees. The initial cost rises slightly, but this is cheaper than refitting each time a use is added.
The first step is a pilot of limited scope — for example, the routes to one primary school, or a limited residential district.
← Previous: 7. If a Government Tried to Misuse It | Next: 9. What It Means for Stakeholders →
FMPS + VRAIO: Public-safety infrastructure built on streetlights, and the governance of AI outputs | Yusaku Fujii (Gunma University) and NPO e-JIKEI Network Promotion Institute
This site presents a research-stage concept; all figures are estimates.