Dash cams have traditionally served a relatively simple purpose: recording what happens on the road. In the event of an accident or other incident, the recorded footage could be retrieved and used as evidence. However, advances in disruptive digital technologies like artificial intelligence (AI) and IoT are changing the role of these devices.
Smart dash cams can now do more than continuously record videos. They can analyse footage in real-time to identify potentially dangerous situations, monitor driver behaviour, and generate alerts. When combined with cellular connectivity and cloud platforms, these capabilities allow information to be immediately transmitted to fleet managers or other authorised users. This connectivity is an important part of the IoT architecture behind smart dash cams, allowing cameras to function as connected sensing devices rather than standalone recording equipment.
This blog post explores how AI is expanding dash cams beyond their conventional role of recording driving footage, with applications ranging from road-infrastructure monitoring and automated inspections to real-time hazard detection and data-driven maintenance. If you want to explore more about smart dash cams, read Transforma Insights’ recently published report, Dash Cams: 106 million connected in-vehicle cameras will improve driver efficiency and road safety in 2035.
Dash cams have traditionally been used for two primary purposes: monitoring driver behaviour and recording footage that can provide evidence in the event of an accident.
One of the established applications of AI-enabled dash cams is monitoring driver behaviour. By analysing the driver and the vehicle's surroundings in real time, these systems can identify risks such as distracted driving, drowsiness and tailgating, and provide immediate alerts or in-cab coaching. For example, Cappagh Browne used AI-powered dash cams to identify risky driving behaviours and support targeted driver coaching. The company reported a 95% reduction in dangerous driving behaviour within one week, alongside an 88% reduction in at-fault accident claims within a year.
Dash cams can also provide an objective visual record of events on the road. Footage can help drivers, fleet operators and insurers understand how an accident occurred, verify incidents and support the assessment of insurance claims. Case in point, RSA Insurance partnered with Lytx to trial its Surfsight AI-12 dash cams with an HGV fleet. Over the 12-month trial, claims frequency fell from 29% to 8%, loss ratio from 48% to 15%, and average claims cost per vehicle-year from GBP1,713 to GBP423 (USD2277 to USD562).
Advances in AI, computer vision, and connectivity are allowing vehicle-mounted cameras to be used for applications beyond driver monitoring and accident evidence. The same cameras can now analyse the external road environment, generate information for road management, contribute visual data to connected and autonomous vehicle systems, and monitor activity inside the vehicle cabin.
AI-enabled dash cams can turn vehicles into mobile sensing platforms that continuously collect information about the roads and surroundings as they travel. Instead of relying entirely on dedicated inspection vehicles or manual surveys, road authorities can use cameras mounted on vehicles already travelling across the road network. AI can analyse the captured footage to identify potholes, cracks, damaged road markings, waterlogging and other infrastructure problems, while location information can be used to map the detected issues.
In Cangzhou, Hebei, an AI highway maintenance system uses vehicle-mounted cameras to identify 14 categories of road defects and safety events. Since its trial began on 30 November 2024, it has covered 531 km of roads per day, identified more than 8,800 defects and reportedly increased inspection efficiency by nearly 60%.
In March 2026, the National Highways Authority of India (NHAI) announced AI-powered Dashcam Analytics Services across approximately 40,000 km of national highways. Cameras mounted on Route Patrol Vehicles identify more than 30 types of defects and anomalies, with the data integrated into NHAI's central Data Lake for road-condition and maintenance monitoring.
In November 2025, the Hawai‘i Department of Transportation (HDOT) began distributing 1,000 dashboard cameras to residents under its 'Eyes on the Road' campaign. Using Blyncsy's AI, the cameras analyse road infrastructure including guardrails, pavement markings and road signs, with footage uploaded to the cloud through cellular connectivity for analysis by maintenance teams.
Mitsui Sumitomo Insurance and UrbanX Technologies launched Dashcam Road Manager in 2021, combining dashcam imagery with AI to detect and map road damage and assess individual road sections. The service was implemented across eight prefectures and 14 municipalities to support road inspections and repair planning.
Surrey County Council uses computer-vision cameras mounted on highways vehicles to automatically identify potholes as inspectors drive, reducing reliance on manual road inspections.
Signal Lab at VŠB - Technical University of Ostrava developed an AI system that analyses conventional dashcam footage to identify potholes, drainage channels and sewer grates. In AI-MATTERS testing, the system detected 45 of 50 defects (90%), compared with 88% for trained human inspectors.
The use of vehicle cameras is also expanding into the development of connected and autonomous vehicles. Here, the camera's role is not primarily to record evidence or identify driver behaviour. Instead, visual information can become an input into systems that need to understand the vehicle's surroundings. AI-powered camera systems can identify objects, road markings, pedestrians, vehicles and other elements of the driving environment. Case in point, Toyota Research Institute-Advanced Development (TRI-AD) and CARMERA used commercially available dashcam drive recorders mounted in Toyota test vehicles to collect road imagery in Tokyo and generate HD maps for automated driving. The resulting maps detected key navigation features such as lane markings, traffic signals and signs, with the project achieving 40 centimetres relative accuracy.
Dash cams and other vehicle-mounted cameras can also generate large volumes of real-world driving data. This data can be used in the development, testing and validation of computer-vision and autonomous-driving systems, helping developers understand how vehicles and perception systems perform across different road and traffic conditions. For example, Nexar uses connected dashcams to collect real-world, including events such as collisions, which can be used to train and benchmark autonomous-driving systems.
AI is also extending camera-based monitoring from the road into the vehicle cabin. Cameras can analyse activity inside the vehicle and identify events involving drivers and passengers. Cabin-monitoring applications can include detecting passenger occupancy, monitoring seat-belt use, identifying unusual passenger behaviour and detecting occupants or objects left inside a vehicle. In commercial and public transport, these capabilities can also support safety and security monitoring. Case in point, Samsung Electronics installed AIMatics' Roadscope 9 AI camera system across more than 2,000 commuter buses. The system includes a cabin camera for internal incident recording, alongside cameras monitoring the road, driver and vehicle blind spots. It provides fleet managers with visibility into cabin activity alongside driver and road monitoring, demonstrating how AI-enabled vehicle cameras can extend monitoring beyond the driver and road environment. The solution resulted in a 41% lesser red-light violation and a 71% reduction in inattentive driving incidents.
The discussion in this blog suggests that the role of AI-enabled dash cams is likely to extend well beyond recording driving footage or monitoring driver behaviour. As AI becomes better at interpreting visual data, dash cams could increasingly act as mobile sensing platforms, continuously collecting information about vehicles, roads and surrounding infrastructure. The combination of AI, IoT connectivity, and cloud platforms could also make this data more useful beyond the vehicle itself. Information captured during everyday journeys could be better analysed, shared with authorised users and integrated into wider systems for road monitoring, maintenance, and safety.
