How AI is supposed to increase safety in trams - Bremen's Trams Deploy AI to Detect Aggression in Real Time

Bremen's public transport system is testing AI technology to boost safety on its trams. The system, introduced in April, uses cameras and real-time analysis to detect conflicts and aggressive behaviour. Fifteen trams have already been fitted with the new tools as part of a nationwide pilot.

The AI works by monitoring passenger interactions through onboard cameras. It first pixelates faces to ensure anonymity before scanning for signs of trouble. If it spots rowdy or violent behaviour, the system alerts drivers and control centres immediately.

Staff can then review the footage and decide whether to intervene or call emergency services. The technology has been approved by Bremen's data protection commissioner, ensuring compliance with privacy laws.

Each year, Bremen's trams see around 150 disputes, including 90 verbal attacks and 30 to 40 physical assaults. The AI aims to reduce these incidents by providing faster responses. By the end of 2024, the city plans to extend the Gemini system to a third of its tram fleet.

Eventually, every tram and bus in Bremen will use the AI safety measures. The pilot began in April, but specific figures on its early impact remain unavailable.

The AI system marks a step toward safer public transport in Bremen. With real-time monitoring and privacy safeguards, it offers a way to address conflicts before they escalate. The city expects to roll out the technology across its entire fleet in the coming years.