A new AI model is taking aim at a question most drivers don’t ask soon enough. How likely are you to crash before you even start the engine?
The system looks at how you behave behind the wheel, pulling in signals like eye movement, heart rate, and personality traits to flag warning patterns early. Instead of waiting for real-world mistakes, it relies on simulated driving tests to surface behaviors linked to dangerous outcomes.
Early results suggest it can distinguish safer motorists from those more prone to serious errors. That could make it valuable in sectors where safety carries real consequences, including delivery networks and commercial transport.
During testing, participants are placed in a controlled virtual driving setup where attention, reaction time, and stress levels are monitored continuously.
Eye tracking shows where drivers focus and how long they stay locked in, helping reveal lapses in attention or slower responses. At the same time, heart rate data reflects cognitive strain, which can shape how decisions are made under pressure.
The model also factors in personality traits that influence risk tolerance and control. Together, these inputs give a more layered view of driver behavior, going beyond simple mistake tracking to identify patterns tied to higher crash likelihood.
For fleet operators, the use case is immediate. Screening candidates based on behavioral signals could help reduce accidents, lower insurance exposure, and limit operational disruption.
Rather than relying only on driving records or basic evaluations, companies could filter candidates before they’re hired. That shifts safety efforts earlier in the process, especially for roles where a single error can have serious impact.
There are tradeoffs to consider. Using biometric and personality data in hiring raises privacy and fairness concerns, and simulator-based signals won’t always reflect real-world conditions.
The model is still being validated in controlled settings, which leaves an open question about how well results carry over to real roads. Driving outside the lab introduces unpredictability that simulations can’t fully capture.
Next steps will likely involve testing with real drivers across a wider range of environments. That will show whether signals like gaze patterns and stress responses stay consistent when conditions change.
If those results hold, adoption in commercial fleets could follow quickly since screening systems are already in place. For everyday drivers, any move into licensing or insurance will depend on regulation and how comfortable people are with this level of analysis.
The bigger shift is already clear. Driving risk may soon be assessed before you even turn the key, and that could reshape how safety is managed from the start. If this holds up, accidents may stop looking random and start looking preventable.
BYD has launched the third-generation Qin Plus DM-i in China, and the first thing that caught my eye is its price. According to CarNewsChina, the plug-in hybrid sedan starts at 79,900 yuan, which is about $11,980. That’s an incredibly low price for a brand-new car, especially by US standards.
What’s even more impressive is the range. The Qin Plus DM-i can drive up to 137 miles on electricity alone, which should cover most people’s daily commutes without using a drop of gas. Once the battery runs out, the 1.5-liter engine takes over. Thanks to its 17-gallon tank and roughly 91 mpg with a drained battery, the combined range reaches up to 1,423 miles.
Apple won the PC market’s worst quarter, but that’s not the entire story. It won the race by stumbling less than everyone else.
PC makers shipped just 57 million computers worldwide in Q3, which might look good, but it’s a 21% year-over-year plunge (via Counterpoint Research). And despite Apple’s Mac shipments falling by 11%, to 5.6 million, its market share rose from 8.7% to 9.8%. How did that happen?
Update (October 22 23:54 PT): In response to our query, GM clarified that offline recording support is already available on eligible vehicles. Like Tesla's Sentry Mode, it requires owners to connect a USB drive to save incident recordings, which can be viewed on the vehicle's infotainment screen. The upcoming Connected Cameras feature will build on this by giving owners remote access to these recordings via the revamped OnStar app, allowing them to check footage without walking out to their car.
General Motors has announced plans to add a new Connected Cameras feature to its OnStar service starting next year, giving drivers a way to monitor parked vehicles remotely and capture footage of certain incidents.