However, the trial also revealed errors in the AI system, such as misidentifying children as fare dodgers and failing to differentiate between different types of bicycles. Privacy experts have raised concerns about the accuracy of the system and the potential for expansion to include more sophisticated detection systems or facial recognition software. Despite these concerns, TfL plans to expand its use of AI to detect fare dodging at more stations across London.
Key takeaways:
- Transport for London (TfL) used AI surveillance software to monitor the movements, behavior, and body language of people using the London Underground to detect crimes or unsafe situations.
- The AI system was tested at Willesden Green Tube station from October 2022 to September 2023, generating over 44,000 alerts, including potential fare dodging, aggressive behavior, and safety incidents.
- Privacy experts have raised concerns about the accuracy of the AI system and the potential for such surveillance systems to be expanded to include more sophisticated detection systems or face recognition software.
- Despite the AI system's ability to detect certain behaviors, it also made errors, such as flagging children as potential fare dodgers and failing to differentiate between different types of bikes.