{
  "id": 11401280,
  "title": "How accurate is Bengaluru’s AI traffic enforcement? | Explained",
  "url": "https://urgent.news/2026/10/02/how-accurate-is-bengalurus-ai-traffic-enforcement-explained-11401280",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-02T08:59:30.000Z",
  "source": {
    "name": "The Hindu",
    "slug": "the-hindu",
    "url": "https://www.thehindu.com/sci-tech/technology/how-accurate-is-bengalurus-ai-traffic-enforcement-explained/article71532005.ece"
  },
  "original_language": "en",
  "account": "In mid-September, a Bengaluru motorist received an online traffic violation notice despite having no pillion rider on his scooter. The Bengaluru Traffic Police (BTP) attributed the error to their Artificial Intelligence (AI)-based violation flagging system, which incorrectly identified a guitar strapped to the motorist's back as a human. The matter was rectified through a manual review.\n\nThe BTP employs three systems to issue contactless challans: Intelligent Traffic Management System (ITMS), Field Traffic Violation and Regulation Analytics (FTVR), and Public Eye, which allows citizens to upload pictures anonymously. The ITMS, introduced in December 2022, uses AI and machine learning to automatically detect traffic violations, sending the data to the Traffic Management Centre (TMC) for further review. According to Joint Commissioner of Police (Traffic), Karthik Reddy, the AI system boasts nearly 99% accuracy, with minimal cases of wrong flagging.\n\nHowever, an investigation by The Hindu revealed that the 0.13% daily rate of challenged challans cannot accurately measure AI accuracy. After reviewing the data, Reddy claimed that the AI accuracy could drop to around 90% after manual review of violations. Further analysis suggests a drop to 80% by the end of 2023, particularly in violations like pillion riding without a helmet, seatbelt violations, and signal jumping.\n\nDespite the machine learning and data feeding improvements, Reddy emphasized the necessity of human intervention to verify certain aspects. He pointed out that at certain junctions, zebra crossing lines are faded, leading the AI to flag motorists unnecessarily. The BTP acknowledged that such inconsistencies require human oversight, as managing civic bodies' shortcomings is beyond their control.\n\nFor instance, the AI may flag motorists who take off their helmets for prolonged periods waiting for traffic signals to turn green, even though they are wearing them. Similarly, the AI has mistakenly identified right turns as jumps, especially in areas where right turns are paused briefly while straight driving is permitted. Additionally, individuals wearing black shirts, which resemble seat belts in color frequency, have been incorrectly flagged for not wearing seat belts.\n\nThe BTP aims to enhance the system's accuracy while acknowledging its limitations.",
  "summary": "Amid debates around the shortcomings of using AI in smart policing and rule enforcement, The Hindu looks at how the Bengaluru Traffic Police’s Intelligent Traffic Management System works, how accurate are the violations flagged, how are the challans generated and the means through which erroneous cases can be challenged.",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 2,
    "also_reported_by": [
      {
        "outlet": "The Hindu - Sci-Tech",
        "title": "How accurate is Bengaluru’s AI traffic enforcement? | Explained",
        "url": "https://urgent.news/2026/10/02/how-accurate-is-bengalurus-ai-traffic-enforcement-explained",
        "published": "2026-10-02T08:38:10.000Z"
      }
    ]
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}