AI for Road Safety: Saving Lives Through Intelligent Systems

Key Topics Covered

• National Road Safety Data Infrastructure: IRAD/EDAR database system and data-driven governance models
• AI Applications in Road Safety: Electronic monitoring, enforcement, predictive analytics, and V2V (vehicle-to-vehicle) communication
• Policy Frameworks: Motor Vehicles Act amendments, data-sharing policies, and emerging regulations
• Youth Engagement Programs: Saraka Mitra volunteer program, Thini AI platform, and National Road Safety Hackathon 2026
• Underreporting & Data Gaps: Challenges in capturing unreported accidents and hyperlocal interventions
• Vehicle Safety Focus Areas: Two-wheeler safety (45% of fatalities), pedestrian safety (20% of fatalities), and vehicle engineering
• Pre-Licensing Education: Integration of driving skills into school curricula and AI-enhanced learning systems
• Empathy & Behavioral Change: Emphasis on human-centered design and cultural approaches to road safety awareness

---

Key Points & Insights

1. National Database as Foundation: India's IRAD/EDAR system attempts something "nowhere in the world attempted"—centralizing crash data, police station information, and population data across 140 crore citizens and 18+ police station jurisdictions into a single, unified database to enable evidence-based policy.

2. Process-Centric Over People-Centric Governance: The shift is from individual decision-makers to systematic, data-driven processes. Governance is "moving from people-centric to process-centric," reducing bias and enabling scalable interventions across districts and states.

3. Data-to-Enforcement Gap: Despite abundant camera data on high-speed corridors, the primary bottleneck is not data collection but data analysis and enforcement without human intervention. AI must bridge the gap between "capturing data" and "providing evidence for legal action."

4. Severe Underreporting Problem: Example: Bihar registers 100 accidents with 95 deaths (95% fatality rate), far exceeding the national average of 36%, indicating massive underreporting of minor accidents. This invisible data is critical for prevention but absent from datasets.

5. Two-Wheeler Safety as Neglected Domain: 45% of road fatalities are two-wheeler related, 20% are pedestrian deaths, yet industry design is car-centric. V2V communication and vehicle feedback systems are being mandated to address low-speed, two-wheeler-specific accidents (35,000 deaths in two-wheeler collisions annually).

6. Vehicle-to-Vehicle (V2V) & Vehicle-to-Infrastructure (V2I) Mandate: The government has mandated V2V and V2D communication with allocated 30 MHz frequency domain, enabling real-time collision avoidance and behavioral feedback before accidents occur—moving from reactive enforcement to predictive prevention.

7. Pre-Licensing as Life Skill Education: Driving skills are not taught in schools despite computers being mandated. Proposal: integrate driving education into school curricula at age 16, making it a foundational life skill for all (particularly girls), which increases road awareness even for non-drivers.

8. Thini Platform: An AI-enabled learning environment (beta) designed as a three-gate system: knowledge gate, skill gate, and practice gate. Named after traditional Indian courtyard learning spaces, it combines cultural pedagogy with AI-driven personalization for safer driving behavior.

9. Hyperlocal Intervention & Youth Volunteerism: Saraka Mitra program recruits Gen-Z volunteers (15-day training) as bottom-up road safety auditors in 100 high-fatality districts, recognizing that centralized policies often fail at hyperlocal, rural, and newly motorized communities.

10. Empathy as Core Principle: Repeated emphasis on human empathy ("factor the human whichever role you are in") counterbalances technical solutions—the belief that awareness and behavioral change, not just technology, prevent accidents. Example: girls avoiding helmets to "not crumple hair" reflects education gaps, not enforcement gaps.

---

Notable Quotes or Statements

• Professor Wenitesh Bala Subramanyan: "When you make it idiot proof, you make better idiots." — On the importance of designing intelligent systems rather than restrictive ones.

• Professor Wenitesh Bala Subramanyan: "Empathy. Please factor the human, whichever role you are in—whether you're on a road, whether you're making a device, whether you are going to apply it—have an empathy towards everybody."

• RP Shukla (Director, Road Safety Cell): "Road safety is not a thing which is a social problem. It is a personal also... Mobility consumption is also a factor."

• RP Shukla: "The main problem is speeding and traffic rules violation. AI can provide evidence before and without human intervention."

• Pankai Jarwal (Chief Engineer, MRTH): On helmet compliance: "Dear, you need the head for hair. So put it."—Illustrating the gap between legislation and behavioral understanding.

• Pankai Jarwal: "Roads are for you, not for challans. Use them judiciously."—Reframing road safety as a rights issue, not a revenue-generation issue.

• Atul Singh (COO, IIT Madras): Announcing National Road Safety Hackathon 2026 with first 50 registrations receiving IIT Madras mentorship—positioning youth as solution-builders, not just problem-recipients.

---

Speakers & Organizations Mentioned

Key Speakers:

• Professor Wenitesh Bala Subramanyan — IIT Madras, Center of Excellence for Road Safety (primary speaker)
• Atul Singh — COO, Center of Excellence for Road Safety, IIT Madras
• Shri RP Shukla — Director, Road Safety Cell, Ministry of Road Transport & Highways (Government of India)
• Shri Pankai Jarwal — Chief Engineer, Ministry of Road Transport & Highways
• Deepak Raja — Representative, Volvo Group
• Dr. Pushkar Patwaran — Co-founder, Light Metrics Technologies
• Himmani Suri — Moderator/Organizer, Center of Excellence for Road Safety

Organizations & Institutions:

• IIT Madras (Indian Institute of Technology Madras) — Lead research and implementation institution
• Ministry of Road Transport & Highways (MoRTH) — Government policy body; initiated national database
• National Informatics Center (NIC) — Implementation agency for IRAD/EDAR
• Volvo Group — Industry partner in road safety technology
• Light Metrics Technologies — AI/tech startup working on road safety
• Ministry of Youth Affairs & Sports — Collaborator on Saraka Mitra program
• Odisha State Government — First state to implement connected, people-centric road safety governance (Rut platform)

---

Technical Concepts & Resources

Databases & Infrastructure:

• IRAD (Indian Road Accident Database) — National unified crash database
• EDAR — Sister/daughter agency system under MoRTH
• Rut Platform — Odisha's people-centric road safety governance system integrating policy planning, implementation, and monitoring

AI & Technology Systems:

• V2V (Vehicle-to-Vehicle) Communication — Mandated real-time collision avoidance and behavioral feedback; allocated 30 MHz frequency domain
• V2D (Vehicle-to-Device) — Vehicle-to-infrastructure communication for hazard alerts
• V2I (Vehicle-to-Infrastructure) — Broader infrastructure-integrated communication systems
• Electronic Monitoring & Enforcement — AI-powered camera systems on high-speed corridors; automated chalan (traffic violation) issuance
• AI-Based Analytics — Data crunching for decision support; automated anomaly detection in accident patterns

Learning & Engagement Platforms:

• Thini Platform (Beta) — AI-enabled road safety learning environment with three-gate model:
  • Knowledge Gate: Understanding road safety principles and etiquette
  • Skill Gate: Practical driving competencies
  • Practice Gate: Real-world application and behavioral change
  • Named after traditional Indian courtyard learning spaces (small front courtyards where elders taught younger community members)

Volunteer & Training Programs:

• Saraka Mitra Program — Youth volunteer program in collaboration with Ministry of Youth Affairs & Sports
  • Target: 100 highest-fatality districts
  • Training: 15 days total (5 days online, 10 days offline)
  • Outcome: Road Safety Auditor certification (co-branded with IIT Madras)
  • Scope: Bottom-up hyperlocal interventions, district road safety committees support

Research Initiatives:

• Data-Driven Hyperlocal Intervention Program — Field perception surveys to capture underreported accident data
• National Road Safety Hackathon 2026 — Open innovation challenge for AI/digital solutions; first 50 registrations receive IIT Madras mentorship
• Road Safety Auditing — Newly institutionalized discipline for analyzing road design, traffic patterns, and intervention points

Policy Instruments:

• Motor Vehicles Act (Recent Amendment) — Mandates electronic monitoring, V2V/V2D communication, and data-sharing policies
• Data-Sharing Policy — Government initiative to enable inter-agency collaboration
• Pre-Licensing Integration into School Curriculum — Proposed mandate to teach driving as a foundational life skill at age 16

Methodological Concepts:

• Process-Centric vs. People-Centric Governance — Shift from individual decision-makers to systematic, data-driven, repeatable processes
• Hyperlocal Intervention — Recognition that top-down centralized policies often fail in rural, newly motorized, and context-specific communities
• Underreporting Analysis — Statistical methods to infer unreported accidents from anomalous fatality-to-accident ratios (e.g., Bihar example: 95% fatality rate vs. 36% national average indicates ~96 unreported minor accidents per 100 recorded accidents)
• Cultural Pedagogy Integration — Using traditional Indian learning models (Thini courtyards) as foundation for AI-enabled education systems

---

Limitations & Caveats

• Audio Quality Issues: The transcript contains numerous repetitions and unclear passages (e.g., "Please slide please. slide please."), making some technical details difficult to extract with certainty.
• Implementation Status: Many initiatives mentioned are still in beta or pilot phases (Thini platform, Saraka Mitra in 100 districts, not pan-India). Scalability and long-term impact remain unvalidated.
• Generalizability: While the talk emphasizes India's global south context, applicability to other regions is not discussed in depth.
• Quantitative Results: The talk emphasizes problem statements and initiatives more than concrete impact metrics, outcomes, or comparative effectiveness studies.