Every year, more and more of us are using AI agents, such as voice assistants, within our cars on a regular basis. In fact, one large-scale survey found that nearly half of consumers today use in-car voice assistants – and by 2022, they predict that approximately 73% of consumers will use an in-car assistant. Up until now, we’ve primarily used these assistants in a utilitarian, command-based format – they wait for us to give them a command, and they carry out our requests accordingly.
What if they also interacted with us proactively, interjecting and making suggestions based on our behavior, surroundings, routines, and preferences? This is where we’re headed. Soon, in-car assistants will evolve into sophisticated, context-aware digital companion agents that engage with us both proactively and reactively, anticipating our needs to ultimately provide us with an enhanced, seamless in-car experience.
But do drivers really want to be “interrupted” by an in-car agent? Would they welcome these proactive interjections, or find them to be a nuisance? In 2019, academic research was conducted on this very topic. The study, entitled Is Now A Good Time? An Empirical Study of Vehicle-Driver Communication Timing, observed 63 different drivers as they drove through a 50-minute simulated course. At various moments throughout their drive, an in-car agent proactively asked them if it was a good time to receive non-driving related information.
Out of the total 2,734 times participants were asked “Is now a good time?” 2,1430 of the responses (~78%) were “Yes.” Perhaps not surprisingly, respondents were most likely to say yes during long, straight stretches, as well as moments where they were stopped (i.e. at stoplights). Less ideal times for the agent to intervene were when the user was trying or about to do a task that required their cognitive attention and/or focus (i.e. turning, merging, or figuring something out like their GPS).
All in all, it indicates quite clearly that people do in fact welcome the notion of an in-car agent proactively interacting with them – so long as it’s at the right moments throughout their journey. Thus, context-awareness is crucial – the agent must be able to sense and understand the driver’s state of mind and preferences, as well as what’s happening inside and outside of the car, to determine the opportune moment to interject. So when and how exactly will it be most beneficial for the agent to use proactivity? Here are 5 prime examples we’ll likely see in the very near future.
By proactively offering up guidance, suggestions, and information, the in-car agent would provide new and/or nervous drivers with some much-needed clarity and peace of mind. As we hop in the vehicle, our in-car agent could remind us to adjust our seating and mirrors, then explain exactly what we can expect in terms of upcoming traffic and weather. If we’re typically nervous about changing lanes, our in-car agent could suggest an optimal time for us to do so. When we’re too intimidated to try out new or ADAS features like self-parking, the in-car agent could help coax us through it – and the list goes on and on.
In the future, as our cars become highly automated, we’ll naturally start to feel a loss of control – especially when we don’t understand what our cars are doing. By sensing our emotions (i.e. when we’re confused, overwhelmed, or distracted) as well as our surroundings, the in-car agent would be able to understand and predict our behavior, carry out actions for us as needed (i.e. hitting the brakes or performing a lane clearance) – and clearly communicate the logic behind its actions to us immediately thereafter.
Let’s say we’re casually turning right at a stop sign, not paying much attention. Out of nowhere, an eScooter juts out in front of us. Our car suddenly slams on the brakes for us, leaving us confused and frazzled. As we begin to process what just happened and recover from the shock, our in-car agent immediately explains why it stopped, and that everything’s under control – giving us a clear understanding of what happened, and helping instill our trust and set expectations with us for future situations of this nature.
Image Source: Coughlin, Reimer, Mehler; MIT/IEEE Pervasive Computing
Another serious problem behind the wheel? Our mood, emotions, and irrational behavior. When we’re tired, distracted, overwhelmed, or angry, it can easily lead to disaster on the road. By proactively speaking with us, the agent would be able to keep us at an optimal driving range, so we’re alert, yet calm. For example, after sensing that we’re not focused on the road, our agent could remind us to pay attention. On the other hand, after sensing that we’re overwhelmed or getting angry in traffic, it could help us destress and calm down.
Another example – let’s say it’s 10 pm, and we’re driving home after a long evening out. Based on our facial expressions (yawning, heavy eyelids, etc.) and the time of day, the in-car agent would detect that we’re tired, and make suggestions accordingly – whether it’s asking if we’d like to call a loved one to chat, or turning on our favorite upbeat music – to keep us more stimulated and awake until we arrive at home safely.
Though this is less important for solo drivers, it would be highly beneficial for professional drivers or those with children in the back seat. Driving can be stressful enough as is, without us having to worry about keeping others amused. Whether it’s the daily commute to and from school, a family road trip, or a long Lyft ride, a proactive in-car agent could ensure that all passengers were optimally occupied, providing them with a more engaging, informative in-car experience.
This would inherently allow the driver to fully focus on the road – so it’s a total win-win. For example, as a taxi driver picks up travelers from the local airport, the agent could ask if they’d like an interactive tour to learn more about the local landmarks as they drive through town. During a family road trip, the agent could help parents keep their children entertained by proactively suggesting interactive games, activities, music, videos, and more.
Have you ever actually sat down and read through your car’s entire manual? Didn’t think so. We all lead busy, hectic lives, and though most of us want to get more out of our cars – whether it’s to maximize our time spent inside of them, to better understand their maintenance needs, or to get the most bang for our buck – we don’t necessarily have enough time on our hands to do so. And as our cars become increasingly automated, with new features constantly emerging, it’s even more difficult to comprehend what’s going on.
Through proactivity, the agent has the unique opportunity to teach us about our car’s new features (or ones we never knew existed) in a way that’s dynamic and effective. It could inform us about a new feature or software update, then talk us through how to use it – answering any questions that may arise as we try it out. It would also take the confusion and friction out of maintenance – proactively reminding us about upcoming maintenance needs or to get our car ready for the winter, then helping us schedule a trip to an authorized garage. This way, we can truly get the most out of our cars – and in turn, we’ll become a more satisfied, confident car owners.
The technology embedded within our cars is constantly advancing – from increased automation, to OTA features, and everything in-between – and it will continue to evolve even further in the years to come. A proactive, context-aware digital companion agent can actively anticipate drivers’ and in-cabin occupants’ needs, so that they can get the most out of their car, and feel more safe, confident, and in-control in the presence of automation. We’re excited to see just how far they’ll take us, and we look forward to the next generation of in-car experience.