Road crashes kill nearly 1.2 million people every year, yet there is no universal way to measure driver risk across global fleets.
The Fédération Internationale de l’Automobile (FIA), the global governing body for motorsport and the federation for mobility organisations worldwide, has developed an AI-powered Driver Safety Index (FIA DSI) to change that reality, aiming to create a standardised risk score that can be compared across countries, fleets and telematics providers.
Launched at the FIA Asia Pacific Congress in Thailand last year, the FIA DSI is now moving into commercial deployment. A number of organisations have already been trialling the platform anonymously, with the FIA set to reveal its first customers this autumn as it finalises agreements with companies in the logistics, energy, oil and municipal services sectors.
For multinational fleets, comparing driver safety can be difficult because telematics providers each use their own scoring systems. The FIA DSI sits above those existing systems as a neutral layer, standardising the data so a risk score generated in one country carries the same meaning as a score in another.
“You can’t improve what you can’t measure consistently,” Tim Shearman, FIA Deputy President for Automobile Mobility and Tourism, tells Zag Daily.
“By creating a universal language for assessing driver risk, the FIA Driver Safety Index fills that gap with an independent benchmark that enables organisations to identify where action is needed most and take proactive steps to reduce the risk of being involved in a crash.”
How does the DSI work?
The system is powered by AI technology from Greater Than, a Swedish company specialising in risk intelligence derived from driving data.
GPS data is collected once a driver’s journey begins and passed through an application programming interface (API), the connection that lets the fleet’s telematics system feed data automatically into the FIA’s platform. Each trip is analysed continuously and compared against a database Greater Than says contains more than seven billion driving patterns gathered over two decades from 106 countries.
That comparison produces a numerical risk score, with a baseline set at 1,000. Scores below the baseline indicate a driver, fleet or geography carries a higher likelihood of being involved in a crash, while higher scores indicate lower risk. Fleet managers can use the scores to compare safety performance by country, region, industry or vehicle type and to identify where training or policy interventions are most needed.
Organisations that want to go further can add Insights Plus, a deeper layer that looks at individual driver behaviour and crash likelihood rather than fleet-wide trends.
Not to be confused with the FIA Road Safety Index
The FIA DSI is separate from the FIA’s existing Road Safety Index, an assessment-based rating scheme already used by companies including robotaxi maker Waymo and truckmaker Scania.
Where the FIA Road Safety Index examines an organisation’s governance, processes and overall safety culture, awarding a rating similar to an ISO certification, the DSI measures actual driver behaviour using telematics data.
“The two Indexes are designed to complement one another,” Shearman says. “Used together, they give organisations a complete system to assess road safety performance from strategy and policies through to day-to-day operations.”
He adds that combining an independent assessment for road safety management and culture with real-world driver risk insights creates a framework for continuous improvements.
Stress-testing the algorithm
Fleet operators are increasingly expected to report independent scoring in their environmental, social and governance (ESG) reporting, so the FIA required an independent validation of the underlying AI.
The review was carried out by Anders Arpteg, Ph.D of Artificial Intelligence. Greater Than commissioned the work but it was conducted independently of both the AI company and the FIA.
Arpteg’s review centred on whether the model was genuinely picking up driver behaviour rather than simply learning to spot patterns tied to time or place. The AI was tested by stripping out all time, date and location data and it was able to flag known high-risk periods, such as late nights at weekends, purely from movement data.
“By analysing pure movement patterns alone, the model successfully detects the intrinsic behavioural signatures of elevated risk,” Arpteg says. “This ‘blind’ correlation to real-world seasonality proves that the AI is not simply following a clock or a map but is sensing the physical reality of driver behaviour.”
He also distinguished the model from conventional telematics which typically infers risk from proxies such as harsh braking, noting that it is instead trained on real-world driving data linked to actual crash outcomes.
Arpteg found particularly strong correlations between the model’s risk scores and fatal or serious crash outcomes across a wide range of countries and confirmed that training and validation data had been kept strictly separate throughout, with exposure weighting removed to avoid skewing the results.
Who is using the DSI?
The FIA has not yet disclosed which organisations will adopt the Index, citing ongoing contract negotiations, but says formal announcements covering companies in logistics, oil, energy and municipal services are expected this year.
One energy company, Equans BeLux, is already trialling the tool, having previously taken part in a dedicated FIA Smart Driving Challenge, with initial results due at the end of August.
Data ownership remains with the fleet operator throughout. Because the underlying GPS data already flows from vehicles to telematics providers under existing commercial arrangements, the FIA’s role is to extend that data sharing to its own database, aggregating and anonymising it before it is used for benchmarking.
“We are not taking ownership of organisations’ data; we are providing a trusted benchmark that allows organisations to measure themselves consistently while helping advance road safety on a much broader scale,” Shearman adds.
Access to the platform is sold on a subscription basis, with monthly, quarterly and annual options depending on how frequently an organisation wants its scores refreshed.
A global index in waiting
The FIA’s ambition extends beyond individual fleet contracts. It wants the FIA DSI to become a publicly available global index of driver risk, in the way ESG ratings or crash-safety ratings have become reference points for entire industries. That would mean opening the system to more sectors and, eventually, more vehicle types.
“Road crashes remain one of the world’s biggest public health challenges and we want to create a common reference point that helps businesses, public authorities and organisations across the mobility sector make better-informed decisions and drive meaningful improvements in road safety,” says Shearman.
Autonomous vehicles are part of that longer-term thinking, with the FIA pointing to its existing relationship with Waymo through the FIA Road Safety Index as a possible bridge into the sector.
