Reduce climate impact: lower your fleet's emissions by 20% Learn more
2 May 2024

Sten Forseke: 20 years of AI, where it all began

Hand holding a lightbulb with the sky in the background. Logo of Greater Than 20 years of AI and profile image of Sten Forseke.

In the days before Greater Than, I founded a company that designed and manufactured electronic information systems and telematics for automotive OEMs, which was later listed on the New York Stock Exchange. The focus was simple: to deliver solutions to improve efficiencies, as well as safety and security, for vehicles, cargo, and drivers. This was successfully achieved.

Traditional telematics parameters of harsh braking, acceleration, cornering, “speed to speed sign” and so on, were used to determine driver safety. And, although it was clear that such events had a clear link with factors such as smooth driving, safety, and vehicle wear-and-tear, my experience with lean production told me it was possible to get more information from the data.

I knew that, in many ways, telematics measurements were based on the assumption that harsh braking or acceleration increased crash risk. Yes, they could increase crash risk, but what if a driver has performed a harsh brake to avoid a crash? Also, what about drivers who adhered to the speed limit, but paid no attention at all to what was happening around them?

The start of something big

Having met Anders Lindelöf late in 2003, I approached him in early 2004 and together we started to consider how different “events” should be weighted. For example, how should a harsh acceleration be valued in comparison to a harsh brake? Which was worse? And, what should the value be if someone traveled way above the speed limit for one minute, compared with someone driving slightly above the limit for 10 minutes?

As an inventor and problem solver, Anders was able to investigate on a deep level the capabilities of the “rules based” model. As he added in more and more criteria to the system – from 30 to approximately 2,000 rules – Anders developed the training algorithm. Using an early form of artificial intelligence (AI) called FASILL (fuzzy logic), Anders was able to identify patterns in driving behavior, rather than isolated events. Over time, he was able to conclude that there was no significant link between driving events, crash risk, and fuel consumption.

Following this, we determined that we were able to use the AI to measure fuel consumption, getting scarily close to reality! And, using the same technique, we started to look at crash probability. The rest, as they say, is history! We set up Greater Than and soon welcomed Liselott Johansson, our CEO, on board. Together, we knew we were in a position to share our data analytics capabilities with the world, to help organizations measure, manage, and mitigate risk and climate impact.

Making a difference, globally

Today, Greater Than works with insurance companies, mobility service providers, driver risk management companies, mobility, and automotive OEMs around the world to enable them to predict crashes before they happen, and to measure the driver factor on climate impact. In other words, “to see the future”.

For insurance companies, the ability to predict who will be involved in a crash – with no need to wait for months of data and its subsequent analysis – means they can accurately price new and existing customers, mitigate risk and improve loss ratio. The real-time crash probability insights provided by Greater Than also enable dynamic pricing models, such as usage-based or behavior-based insurance solutions.

Within telematics and driver risk management, our insights enable the prompt identification of high risk (and low risk) drivers, facilitating timely and cost-effective risk mitigation. Typically, our AI is used as an “extra layer” to the organization’s existing data insights, such as driver behavior, mileage, vehicle tracking, routing and efficiencies. The benefit of the extra insights is to gain crash probability insights and climate impact, as a result of driver attitude.

A spotlight on sustainability

Recently, we’ve seen interest in our climate impact insights skyrocket. This is largely down to the ramping up of Environmental, Social & Governance (ESG) reporting regulations including the Corporate Sustainability Reporting Directive (CSRD). Added to this is the demand from investors, customers, supply chain partners and other stakeholders, for companies to be more transparent about their climate impact.

Our key differentiator has always been, and remains today, that our AI can measure a driver’s influence on crash probability and climate impact regardless of where they are in the world or what vehicle they are driving. This means that, for companies looking to reduce their impact on the planet, we give them valuable and actionable data that enables them to make a rapid difference, in little time, with little effort, with no negative impact on their operations.

For many companies involved in transportation – taxi firms, delivery companies, and service providers are good examples – it’s challenging to reduce CO2 emissions, despite the pressure being on. The default assumption by many organizations is that a reduction in emissions can e.g. be achieved by driving less or switching to more fuel-efficient vehicles. Yet, we have shown that simply increasing drivers’ awareness of the way they drive – increasing their focus, attention, and vehicle control, can reduce CO2 emissions by an average of 20%.

AI like no other

So, how does our AI technology work? Since 2004, we have been analyzing data from 106 countries and 1,600 cities, training our database with real world trips and their outcomes, including claims, no claims and fuel consumption. Today, our database contains over 7 billion driving patterns, meaning it can quickly compare real-time data to existing patterns to look for similarities. In fact, with only 1km of driving data, Greater Than’s AI can start to identify the driver’s influence on crash probability and climate impact.

While traditional telematics roughly analyzes 5% of every hour – maybe 150 data sets – our AI measures 3.6 billion data sets per hour. And, comparing it to an insurance environment, our AI takes the same number of decisions as 37,000 actuaries every day. It does it in the same way every time, with no bias; it measures both positive and negative behavior throughout a whole trip. And the beauty of it is that it’s constantly learning. We’re now getting very smart at predicting specific crash type as well as severity.

The value of our data analytics has been proven many times over the years. I’m proud that, following a 6-year review, the global governing body for motor sports, the Federation Internationale de l’Automobile (FIA), selected our technology to use in its FIA Smart Driving Challenge, a worldwide challenge rewarding smart, safe, and eco-friendly driving. This was down to our AI being comparable globally, independent of vehicle make or model. Due to its effectiveness in driving change, the technology was showcased by the FIA and Greater Than during the United Nations’ Climate Change Conference, COP28.

Added to that, our AI is endorsed by the Swedish Energy Agency, supported by funding from the EU, and awarded by the WWF Climate Solver Award for its potential to influence global CO2 savings using the GHG protocol.

These are recognitions I am extremely proud of, along with other nominations and awards we’ve received over the years. I’m also proud that our AI makes a difference on a daily basis for our customers around the world, helping to change thinking around the way risk is measured, priced, and managed.

Looking to the future of AI

I’m passionate about continuing to push for AI’s predictive capabilities to be adopted on a wider level in the insurance industry, as the industry is key in making a positive difference to road safety and sustainability. If all insurance companies were to price risk correctly – based on actual crash probability – they would reduce road injuries by 10 million per year and fatalities by 300,000 per year. Currently, the drivers least likely to crash finance those with highest crash risk, and there is no incentive for drivers to adopt safer, more sustainable behaviors. AI has the power to change that.

To summarize our AI’s 20-year history, what an incredible journey is has been! Anders and I recognized the potential of technology to determine risk insights before it was even known as “AI”. Going forward, the Greater Than team will continue to train and advance the AI with a vision of it being standard in every vehicle. This will ensure a safer, more sustainable future for us all.