Picture this – you hop into your car, ready for the daily morning drive. But last night you didn’t sleep nearly as soundly as normal. You’re groggy, sporting puffy eyes, and your normal radiant smile is nowhere to be found. As you enter your car, your car’s “affective computing” program analyzes your face and immediately knows how you’re feeling. This registers in the system and the car is now on a higher safety alert as your response time to potential obstructions might be compromised. Oh, and your car puts on soothing music and injects lavender into the air as well.

You’re likely thinking, “my 2010 Accord can do this?” No, your 2010 Accord can’t, but researchers are close to making this a reality in those models that choose to adopt it very soon. Scenarios like this are all part of artificial intelligence (AI) and our car’s ability to get to know us just a little bit better every day that passes.

AI systems can be trained to differentiate between a dog and a cat, for example. In the same way, they can also differentiate between an angry and a happy tone, as well as a frustrated and satisfied expression. Affective computing essentially logs sets of labeled data and then makes correlations. A smile equates to a happy feeling, while shouting is aligned with anger, tears with sadness, and so on. As time goes on affective computing can pick up on tiny, micro-expressions that might appear in response to something but then disappear just as quickly.

Affectiva is a company that originated in the MIT Media Lab. Their algorithms were trained on 5,313,751 facial videos. They’re so robust that they can now distinguish between cultural differences, where a slight smile in Tokyo might mean something different than the same smile in Toronto. This all sounds great if the end-goal is making driving safer or less stressful for the driver. But what about privacy concerns? After all, this is facial recognition, and there’s been lots of discussion, especially in terms of COVID-19, and how far government can reach into someone’s private life in the name of public safety.

European regulations dealing with driver data are far more developed than in the US. Car companies stateside are in general free from laws that would restrict them from using driver data. This is shifting, but it’s been surprisingly slow. Most would agree that affective computing if used responsibly, has the chance to radically alter (for the better) safety conditions. But first things first – your smile is your smile, not a smile being sold to a potentially nefarious third party.