What if autonomous vehicles didn’t just respond to the road, but to the human experience of it? Recently we’ve been discussing new ideas with AI focussed businesses looking at training AI/ML systems using real human data.
For decades, AI driving systems have trained on telemetry: steering angles, acceleration curves, GPS paths, and sensor data. That is to say they have learned to drive by interpreting the data from the car, not from the humans controlling the vehicles. That is to say the algorithms are not learning from the driver’s point of view.
They haven’t learned what it feels like to be human behind the wheel. The mental strain of decision-making, the fleeting moment of distraction, the subtle stress of a sudden swerve. That’s the missing piece. And biometrics such as EEG (brainwaves), eye tracking, heart rate, and GSR (galvanic skin response) are capable of giving AI a window into our inner world.
This isn’t about gadgets or gimmicks. It’s about elevating how machines learn by teaching them not just what we do, but why we do it, how we interpret the world and how we react to it.
Revealing the Invisible
EEG captures the brain’s electrical activity in real time. It reveals cognitive load, stress, alertness, or fatigue. Eye tracking shows where attention is focused, what’s being processed, and what’s being missed. GSR and heart rate can detect when a driver or passenger is uneasy, even before they are consciously aware.
Together, these signals paint a picture of our cognitive and emotional state, something no camera or LIDAR can see. Two drivers might approach the same corner at the same speed. But one is calm and composed. The other is overworked and missing critical cues. Telemetry alone can’t tell the difference. Biometrics can.
Training AI to Think and Feel
When these internal states are fused with external data, AI could improve on, not just mimic human behaviour. It learns the conditions under which those behaviours occur. It understands that a hard brake preceded by a high-stress EEG spike and fixated gaze likely came from panic, not intent. That a lane change preceded by a head turn and calm vitals is confident, not reactive. These connections allow AI to distinguish between expert decisions and anxious ones, helping systems learn what good and safe truly feels like.
It’s no longer enough for AI to drive like us. It must understand us.
Emotionally Intelligent Machines
Driving is deeply emotional. Passengers feel stress during sharp turns, fear during close calls, and relaxation on a smooth cruise. These sensations influence trust, comfort, and perceived safety, all critical factors for mainstream adoption of robo-taxis, drone mobility, and shared autonomy.
Biometric data gives AI the tools to be not just responsive but considerate. A vehicle that notices rising passenger heart rate can slow down. If EEG detects discomfort, the system can ease off the acceleration or change lanes more gently. If the car senses calm and trust, it knows it’s doing something right.
Concept cars like Cadillac’s InnerSpace hint at this future. Sensors read vital signs. Lighting and temperature adjust to soothe stress. But beyond luxury, this is foundational. A vehicle that responds to how you feel is one that earns your confidence. That’s not a nice-to-have. It’s mission critical for adoption.
The New Signals of Trust
Research is showing that fusing biometrics with traditional driving data leads to higher prediction accuracy for driver intentions, safer responses in hazardous scenarios, and better anticipation of manoeuvres. In one study, combining EEG and eye tracking allowed researchers to predict a driver’s decision to turn left or right up to a second before the action happened. Far earlier than any motion-based model could. That’s the edge biometrics offer insight before behaviour.
In another experiment, a passenger’s EEG signals detected perceived danger faster than the AV’s onboard systems could. Their brain reacted instinctively a gut feeling of risk before the algorithms flagged anything. That signal became an early warning. When AI is trained on such data, it gains a sixth sense. A human one.
This ability to read human cues from driver intent to passenger emotion is how AI evolves from being merely reactive to truly anticipatory.
From Labs to Roads
This isn’t blue-sky speculation. Nissan’s Brain to Vehicle project has already shown how EEG can help vehicles assist drivers by predicting movements milliseconds ahead. Driver monitoring systems like those from Tobii are increasingly common. They track gaze to assess awareness and feed that data into safety protocols.
Research programs across Europe and North America are exploring how biometric feedback can optimize ride comfort in AVs. And drone-taxi developers are quietly studying how to use biometrics to help remote operators avoid cognitive overload or to ensure nervous flyers remain calm mid-flight.
We are standing at the edge of a new paradigm in mobility. One where the AI doesn’t just know where the road goes. It knows how you feel about the ride.
A Human Centred Future
When autonomous systems can read the subtle cues of our attention, stress, and intent, they stop being mere machines. They become collaborators. They don’t just replicate the mechanics of driving. They embody the mindset, the emotions, the awareness that make great driving possible.
Biometrics bring empathy to automation. They allow AI to treat the human not as a control variable, but as a core design principle. That means AVs that learn not only to avoid crashes, but to avoid discomfort. Not just how to get from A to B, but how to do it in a way that feels intuitive, respectful, and calming.
It’s time we stop thinking of AI drivers as replacements for humans and start seeing them as partners copilots trained not just on what we do, but how we think, feel, and react. That’s how we build trust. That’s how we make autonomy acceptable, not just operational.
The future of autonomous mobility isn’t just faster, safer, or smarter. It’s more aware.
Of the road. Of the world. Of you.
