Research Reveals Driver Assistance Technologies Operating Like Autopilots as Drivers Participate in Multitasking

Research Reveals Driver Assistance Technologies Operating Like Autopilots as Drivers Participate in Multitasking

Research Reveals Driver Assistance Technologies Operating Like Autopilots as Drivers Participate in Multitasking


### The Tempting Appeal and Concealed Risks of Semi-Automated Driving Technologies

![Enlarge](https://cdn.arstechnica.net/wp-content/uploads/2024/09/GettyImages-1633254374-scaled.jpg)
*Lane-keeping technologies enable hands-free driving.*
*Image Credit: Getty Images*

The vision of fully self-driving cars has intrigued both the public and car manufacturers for more than ten years. The concept of a vehicle that navigates on its own, permitting occupants to unwind, check emails, or even enjoy a movie during their travels, has resulted in substantial financial investments and remarkable progress in vehicular technology. Yet, despite the excitement, we remain a considerable distance from realizing complete autonomous driving. Instead, automakers are progressively introducing **semi-automated driving systems** that manage functions like accelerating, braking, and steering. While these technologies boast enhanced convenience, emerging evidence indicates they may render us **safer** on the road.

### The Growth of Semi-Automation

Semi-automated driving systems, including **Tesla’s Autopilot** and **Volvo’s Pilot Assist**, are engineered to aid drivers in maintaining lane alignment, adjusting vehicle speed, and managing certain steering functions. These systems are frequently promoted as a solution to lessen the strain of driving, especially during lengthy highway journeys. Nevertheless, they are not designed to completely replace human intervention. The driver is still required to remain vigilant and prepared to regain control at any given moment.

In spite of these expectations, **multiple studies** have demonstrated that drivers often develop an excessive reliance on these systems, resulting in hazardous levels of distraction. The latest evidence stems from two studies executed by the **Insurance Institute for Highway Safety (IIHS)** in partnership with the **Center for Transportation and Logistics at MIT’s AgeLab**. These investigations concentrated on the actions of drivers utilizing Tesla’s Autopilot and Volvo’s Pilot Assist, and the outcomes are troubling.

### The Investigations: Tesla and Volvo

#### Volvo Pilot Assist Examination

The initial study comprised **29 participants** operating **Volvo S90 sedans** equipped with Pilot Assist, a semi-automated driving system. None of the participants had prior experience driving the S90, and most were unfamiliar with semi-automated systems. The participants were divided into three groups, each receiving varying levels of training on how to operate the system. They were subsequently instructed to use the vehicle as if it were their own for a period of four weeks.

The examination revealed that drivers were considerably more inclined to participate in **distracting activities**—such as using their smartphones, eating, or even personal grooming—when Pilot Assist was in operation. Over time, as drivers became more accustomed to the system, their complacency grew, leading to increased multitasking. This outcome emphasizes a concerning trend: as drivers grow more comfortable with automation, their attention to the road diminishes.

#### Tesla Autopilot Examination

The second examination involved **14 participants** driving **Tesla Model 3s** outfitted with Autopilot. Similar to the Volvo study, none of the participants were previously acquainted with semi-automated driving systems. During the research, participants drove a collective **12,161 miles (19,571 km)** with Autopilot activated. The researchers tracked how drivers reacted to **attention-related alerts** from the vehicle, triggered when the system sensed insufficient driver attention.

The study documented **3,858 attention-related alerts**, with 98% categorized as the lowest-level reminder to “apply slight turning force to the steering wheel.” Although the frequency of these low-level alerts increased over time, the number of escalated alerts issued when drivers ignored the initial warning decreased. Yet, despite these reminders, drivers continued to engage in **non-driving activities**, suggesting that the alerts were inadequate for ensuring attentive driving.

### The Complacency Dilemma

Both investigations uncovered a shared concern: drivers become **complacent** when utilizing semi-automated systems. As they become more at ease with the technology, they are more inclined to indulge in distracting behaviors, presuming the vehicle will manage everything. This complacency is perilous because these systems are not infallible. They are intended to assist rather than replace drivers, necessitating ongoing human oversight.

The studies also underscored the inadequacies of present driver-monitoring systems. For instance, the Tesla Model 3s used in the study depended on a **torque sensor** on the steering wheel to determine if the driver was attentive. This sensor only detects whether the driver is applying force to the wheel, which can easily be manipulated without guaranteeing the driver is focused on the road. In contrast, newer vehicles are beginning to incorporate **gaze-tracking systems** to observe where the driver is looking, which might provide a more precise measure of attention.

### What Requires Improvement?

The results from these examinations highlight the necessity for **enhanced safeguards** to prevent