Sleeping in a self-driving car? A new survey shows how Americans feel
New research highlights the gap between using autonomous features and fully relying on them.
Autonomous driving is no longer a science fiction concept. Since Tesla introduced its self-driving system, nearly every major automaker has been developing its own version or adapting similar driver-assistance technologies. Fully autonomous driving is still a future for the majority, but assisted driving is already part of everyday life for many drivers. At the same time, robotaxi services are expanding rapidly, with companies like Waymo regularly announcing services in new cities, making fully autonomous rides an everyday reality. But as the technology spreads, an important question remains: how much do people really trust autonomous driving?
Do people trust autonomous systems
While many people are increasingly comfortable with autonomous systems in everyday situations, that trust drops noticeably when it comes to falling asleep behind the wheel. Attitudes toward autonomous driving suggest openness to the technology, but clear limits remain around personal safety.
In a recent survey of more than 1,000 respondents, 41.5 percent said they would never sleep in an autonomous vehicle under any circumstances. 12.7 percent of the people already feel comfortable enough to sleep in a semi-autonomous vehicle today, and 34 percent said they would be willing to nap in a fully autonomous car in the future. This indicates that trust in the technology is growing, but it will still take some time for people to get used to it and to trust it.
Autonomous system vs. real drivers
Driving while drowsy remains a widespread and often underestimated safety issue on the road. For decades, vehicle safety systems have been developed to reduce human error and lower the risk of accidents, from basic driver alerts to advanced assistance features.
Autonomous driving technology takes that goal further by aiming to eliminate one of the most significant risk factors: human error. In the survey, 73.9 percent of respondents said they were more concerned about tired human drivers than autonomous vehicles, compared to 26.1 percent who felt the opposite. Nearly 60 percent also said the biggest benefit of sleeping in a self-driving car would be removing fatigued drivers from the road.
It is still early days for autonomous driving, but as the technology continues to develop, it is expected to match—or eventually surpass—the abilities of human drivers in many situations. As more self-driving vehicles enter everyday use, the overall risk tied to human behavior on the road is likely to decline. In that scenario, human drivers may become the most unpredictable and risky factor in traffic, rather than the technology itself.
The comfortable minority
It’s true that majority of drivers surveyed said they wouldn’t be able to trust autonomous driving while they sleep, however, a notable share is warming up to the idea of sleeping in an autonomous vehicle someday. The survey found that about one-third of people would consider sleeping in a fully autonomous car in the future, and a smaller group would try it in a semi-autonomous one.
Those who were open to the idea described a very narrow set of conditions (like extreme exhaustion and long highway trips) as the only situations in which they would even consider it. Even then, the median speed they felt comfortable sleeping at was just 46 miles per hour. This reveals an underlying tension between theoretical trust and lived comfort.
Even when respondents acknowledged the benefits of automation, they considered more traditional forms of travel safer for sleeping. The study found that more than two-thirds felt comfortable sleeping on a plane, and nearly as many said the same about trains. More than half said they would be able to sleep on public transit. By comparison, only about a third said they would feel safe on the road near autonomous cars while their children were in the vehicle.
The contrast here suggests that being comfortable sleeping while traveling is more about context. People accept sleep more easily when surrounded by infrastructure designed for it or when responsibility is assigned to a trained professional. Cars remain deeply connected to the idea of personal responsibility, and that association makes it difficult for Americans to be comfortable closing their eyes, even if the system can handle the journey.
Public resistance
In general, the most consistent barrier to drivers sleeping in a self-driving car is the feeling of relinquishing control. Participants in the survey described a sense of vulnerability arising from relying entirely on automation while traveling at highway speeds. This fear was about the ability to trust the technology itself, but also the behavior of other drivers.
There were different anxieties highlighted in public resistance, with respondents expressing distrust of the software that guides the vehicle due to the risk of hacking. There was also a sense of unease about crashes caused by unexpected obstacles, and concern about system failures that might happen too quickly to be corrected or detected by automatic systems.
25% of respondents said they would only feel safe if another person monitored the vehicle, which still contradicts the purpose of a fully autonomous self-driving vehicle. These concerns show that the hesitation is rooted in the gap between engineering promises and personal confidence.
We see the potential danger of going to sleep behind the wheel of a self-driving car as presented by @marcusjcarterai, whose car did an unexpected, sudden turn:
The man in the clip is using Tesla’s self-driving feature when the vehicle suddenly and unexpectedly veers left into oncoming traffic. Luckily, he was awake and made a quick correction, but this minor error could have caused a significant accident. It is unclear what triggered the system to turn. At first, it looks like it got frightened by the bins on the side of the road. It is just a small reminder that even if the technology is rapidly evolving, we still need some time before we can 100% trust it.
Trust is building
The findings of this survey indicate some potential for long-term optimism on the topic. A majority of respondents believe sleeping in autonomous vehicles should be legal, and nearly 60% want autonomous features in their next car. Another 60% said they would sleep in a fully autonomous ride-hailing vehicle 10 years from now, because people trust automation more when they imagine a future where the systems have matured, and errors have had more time to be refined.
Even though there’s some optimism regarding autonomous driving in the future, it’s important to keep in mind that most vehicles on the market today are not fully self-driving, even when they include advanced driver-assistance features. Regulators and manufacturers consistently emphasize that these systems still require an attentive human behind the wheel, since the technology is designed to assist rather than replace the driver.
The National Highway Traffic Safety Administration states, “It is critical that the public understands a vital fact about current technologies: all vehicles sold to the public today require a driver to be fully attentive and cognitively engaged in the driving task and the road ahead with the consumer available technologies today. … While these ADAS technologies are improving and enhancing safety, they are not self‑driving. Misusing driver assistance systems by failing to maintain control of the operation of the vehicle at all times can result in serious and even deadly crashes.”
Takeaway
While autonomous driving systems are not yet advanced or sufficiently regulated to hand over full control to computers, the technology continues to move in that direction. As these systems improve and prove they can outperform human drivers, broader legal approval is likely to follow. At that point, sleeping during a commute may no longer feel unrealistic, and more drivers are likely to trust the technology enough to rest while it drives. Until then, driving remains a human responsibility.
The study
The survey gathered responses from 1,008 participants recruited through Amazon Mechanical Turk, with a nearly even gender split and a wide age range spanning from 18 to 83. The average respondent age was 37, reflecting a broad cross-section of adult perspectives. To ensure data quality, an attention-check question was used and responses from participants who failed it were removed. The survey carries a margin of error of plus or minus 3 percent at a 95 percent confidence level. Clear definitions of fully autonomous and semi-autonomous vehicles were provided at the start and remained available throughout the questionnaire.
