Bengaluru Engineer Builds AI System to Detect Potholes and Trace Road Contractors
The app can search through nearly 2,900 government road contracts to connect the damaged stretch with the relevant tender and contractor responsible for the road.
A pothole complaint usually tells authorities where a road is damaged. A system built by Bengaluru-based engineer Gaurav Sen attempts to answer another question: who is responsible for fixing it?
Sen has developed an AI-powered setup that uses a dashcam, GPS and accelerometer to capture road conditions while he drives. The system processes the collected footage and data to detect potholes and record their location.
What makes the setup different is what happens after a pothole is detected. The application can search through nearly 2,900 government road contracts to connect the damaged stretch with the relevant tender and contractor responsible for the road.
This could be particularly useful when a road is still under a contractual warranty. If a defect occurs during the warranty period, the contractor may be obligated to carry out the repair without charging the government additional costs.
Sen demonstrated the system in a video shared on Instagram in collaboration with ChatGPT India. According to him, he built the application using Codex, with the system sending relevant road data to OpenAI for processing.
The setup is designed to handle road defects that may not always be obvious from a moving vehicle. Sen said the system can also identify potholes that could otherwise be mistaken for speed breakers.
In one test during his daily commute, the system reportedly identified 12 potholes. It then generated complaint records containing photographs, location information and details linked to the corresponding road contracts.
The system can reportedly produce such a structured record in around four seconds, including information such as the pothole photograph, location, tender number and the officer responsible for the road.
The idea addresses one of the practical difficulties with civic complaints. Reporting a pothole is relatively straightforward, but finding the correct authority or contractor responsible for the particular road can be more difficult.
By combining everyday driving hardware with AI-based image analysis and government contract data, Sen’s project attempts to turn road damage captured during a normal commute into documented complaints that can be directed to the relevant stakeholders.
The demonstration has also gained attention on social media, where users have pointed to its potential use in improving accountability around road maintenance.