AI for Safer Roads Innovation Challenge

Can AI tell us where speed limits need to be updated to better protect road users?

CHALLENGE OVERVIEW

AI for Safer Roads Innovation Challenge is an initiative that explores how emerging technologies can transform the way road safety is understood and managed. By combining AI, large-scale mobility data, and geospatial analysis, the initiative seeks to uncover hidden patterns of risk and provide actionable insights for safer road systems. It supports the development of new tools and methodologies that can be applied across countries to improve decision-making and accelerate progress toward reducing road deaths and injuries.  

The initiative led by the Asian Development Bank in collaboration with the World Bank Development Impact Group, AI for Good, ITU, and supported by JFPR and HLTF.

 

This challenge asks: How might we use AI and mobility data to determine where speed limits are misaligned with real-world road conditions, supporting evidence-based speed management across Asia and the Pacific? 


We are challenging data scientists, AI specialists, transport engineers, and policy innovators from ADB member countries to develop an analytical model that: 

  • Assesses whether posted speed limits align with Safe System principles 
  • Identifies road segments where limits are inconsistent with road function or vulnerable road user exposure
  • Produces a spatial output, a map-based visualization, highlighting priority segments for review or intervention 
  • Is scalable and replicable across countries in Asia and the Pacific 

 

This is not about measuring whether drivers are speeding. It is about determining whether the current speed limit itself is appropriate for the road.


 

Applications open May 20

 

 

Solution Features

 

The submitted solutions should demonstrate capabilities in the following areas:

SAFE SPEED ASSESSMENT

Develop a methodology to evaluate whether posted speed limits are consistent with road function, operating speeds, and surrounding land use.

RISK IDENTIFICATION

Identify road segments where current speed limits may be exposing pedestrians, cyclists, and powered two-wheeler users to unacceptable risk.

POLICY-READY OUTPUTS

Produce a Speed Safety Score and geospatial visualization that governments can use to prioritize interventions.

 

DATA PROVIDED

 
Participants will receive anonymized, aggregated datasets from national coverage and state-level coverage, including:
 
  • GPS Probe Data: operating speeds, 85th percentile speeds, speed distributions, posted limits, traffic intensity
  • Road Network Data: functional class, urban/rural classification, intersection density, segment length
  • Mapillary Street-Level Imagery: crowdsourced images with ML-identified road features and signs
  • Optional contextual layers: population density, land use, proximity to schools and markets, powered two-wheeler indicators
 
Participants will have access to the datasets after registration and agreeing to NDA.
 

TIMELINE

  • APPLICATION

     

    By 25 June 2026

    Submit your AI model and geospatial analysis by 25 June.

  • REVIEW

     

    July 2026

    Experts review the submissions. 5 top solutions are shortlisted and announced in September 2026.

  • REFINE

     

    September 2026

    Top 5 teams build their visualizations on ADB's GIS platform and implement received feedback.

  • PITCH

     

    October 2026

    Teams present their solutions to a high-level jury.

 

WHAT'S IN IT FOR YOU?