Technology and Autonomous Vehicles

With the potential addition of autonomous cars and trucks to commercial fleets, commercial insureds should reassess coverage under their commercial auto, products liability, and cyber insurance policies. Automation, particularly the move toward fully autonomous vehicles, raises new coverage questions with new risk exposures for insureds and new opportunities for insurers.

This post is the first in a series examining potential coverage issues under different lines of coverage. We begin here with commercial auto insurance policies. Future posts will examine products liability and cyber insurance policies.

ISO Commercial Auto Form

Many commercial insureds’ auto policies incorporate the Business Auto Coverage Form (CA 00 01 10 13) (“Commercial Auto Form”) promulgated by the Insurance Services Office, Inc. (ISO).  The Commercial Auto Form provides coverage for “bodily injury” and “property damage” caused by an “accident” and “resulting from the ownership, maintenance or use of a covered ‘auto.’”  Although numerous cases have interpreted the scope of this coverage in cases involving vehicles other than cars and trucks, such as in accidents involving mopeds, 4x4s, riding lawnmowers, and other vehicles, courts have not yet determined whether this coverage extends to accidents involving autonomous vehicles.

Autonomous Vehicle Classifications

In a sense, today’s cars and trucks all offer some autonomous features. Some features, such as cruise control or self-parking, can operate independently once engaged by a driver. Only highly automated and fully automated vehicles can operate without driver control, however.

In its current operating guidance for autonomous vehicles — Automated Driving Systems 2.0:  A VISION FOR SAFETY, released in September 2017 (Voluntary Guidance) — the U.S. Department of Transportation’s National Highway Traffic Safety Administration (NHTSA) categorized autonomous vehicles using the International Society of Automotive Engineers’ (SAE) classification of autonomous vehicles levels between Level 0 and Level 5:

SAE Automation Levels
Source: NHTSA, Automated Driving Systems 2.0: A Vision for Safety.

NHTSA’s Voluntary Guidance describe vehicles that incorporate SAE automation Levels 3 through 5 as “Automated Driving Systems” (ADSs). Level 3 Conditional Automation ADSs require drivers to monitor and take control when necessary. Level 4 High Automation vehicles can be driven by a human, but are self-driving and do not require a driver. Level 5 Full Automation vehicles can perform all driving functions under all conditions. Level 5 vehicles include those vehicles with driver controls, as well as vehicles without driver controls, for example vehicles that lack a steering wheel or pedals. Level 5 vehicles without driver controls could incorporate seating spaces designed for conversation or work spaces with tables – but not controls that permit human occupants to assume any driving function. Let’s call these Level 5 vehicles “Personal Mobility Units” or “Level 52” (Level 5 Squared) vehicles.

Are Personal Mobility Units Considered “Autos” under the Business Auto Coverage Form?

ISO’s Commercial Auto Coverage Form identifies “autos” by numerical symbols 1-9, and 19.  Each insurance policy’s declarations page will identify the insured “autos” using those symbols.

Symbol 1 includes “Any ‘auto’” while symbols 2-9 narrow the definition of “auto” to exclude certain categories. ISO also offers the Covered Auto Designation Symbol endorsement (CA 99 54 10 13), which allows insurers to designate Symbol 10 “autos” to meet their insureds’ specific needs.

Is a Level 52 vehicle a Symbol 1 auto (“Any ‘auto’”)?  That depends on the definition of “auto.”  The Commercial Auto Form defines an auto as “a land motor vehicle” or “any other land vehicle” governed by state insurance law:

“Auto” means:

  1. A land motor vehicle, “trailer” or semitrailer designed for travel on public roads; or
  2. Any other land vehicle that is subject to a compulsory or financial responsibility law or other motor vehicle insurance law where it is licensed or principally garaged.

But “auto” does not include “mobile equipment.” CA 00 01 10 13, Section V –Definitions B (page 10 of 12).

Turning to part 1 of the “auto” definition, is a Personal Mobility Unit a “land motor vehicle?”

Of course, no court has considered this question. Those courts that have construed this definition in other contexts have considered whether the vehicle included customarily required features, whether the vehicle “resembled” a vehicle designed for use on public roads, and whether the vehicle was required to be registered by state law.

So, a construction vehicle known as a “Georgia Buggy” was not an “auto” because it: (1) did not resemble an auto; (2) was operated from a platform using “hand controls affixed to bicycle­-like handlebars;” (3) lacked key components, such as a steering wheel, a windshield, a separated driving compartment, lights, turn signals, and mirrors; and (4) was not required to be registered by the state. Harlan v. United Fire and Casualty Company, 208 F. Supp. 3d 1168, 1172‑73 (D. Kan. 2016), appeal dismissed, No. 16-3310, 2017 WL 4863142 (10th Cir. Jan. 13, 2017).

A photograph submitted during the litigation surely shows that the buggy was not an auto:

Georgia Buggy Exhibit
Georgia Buggy Exhibit

Of course, a Level 52 vehicle might not “resemble an auto,” would not have any type of hand controls, and would lack some (but not all) key components, such as a steering wheel.

A crop sprayer that collided with a motorcycle was considered an “auto” under the same policy language because it was subject to compulsory insurance and financial responsibility laws, had four wheels, was self-propelled, had headlights, taillights, turn signal, and “other components similar to road‑ready vehicles.” Berkley Regional Specialty Ins. Co. v. Dowling Spray Service, 864 N.W.2d 505, 508 (S.D. 2015). Under this reasoning, a Personal Mobility Unit is an “auto.”

Turning to part 2 of the “auto” definition, is a Personal Mobility Unit “subject to a compulsory or financial responsibility law or other motor vehicle insurance law? Is it licensed? Is it “principally garaged” in any state?

All ADSs are self-propelled and, like other self-propelled vehicles, will be subject to compulsory insurance or financial laws in most states (with certain exceptions). For example, in Delaware, motor vehicles include all “self-propelled” vehicles except farm tractors, elective personal assistive mobility devices (electric two-wheeled vehicles for one person, limited to 15 miles per hour or less), and off‑highway vehicles. Del. Code Ann. tit. 21, § 101. In Michigan, industrial equipment, electric personal assistive mobility devices, electric patrol vehicles, electric carriages, and commercial quadricycles are not “motor vehicles.” Mich. Comp. Laws § 257.33. In California, self‑propelled wheelchairs, motorized tricycles, and motorized quadricycles are not motor vehicles if operated by a person who, by reason of physical disability, is otherwise unable to move about as a pedestrian. Cal. Veh. Code § 415. Under these and other regulatory regimes, all ADSs should be subject to existing compulsory insurance or financial laws in most states (pending any change in regulatory regimes to accommodate ADSs).

No state excludes autonomous vehicles from registration. An auto is generally licensed in the state where it is registered. At the time of registration, the state will issue the appropriate number of license plates, which the owner or registrant is then required to display. See, e.g., Cal. Veh. Code § 6700. A tractor-trailer that displayed Nebraska insurance plates was indisputably licensed in Nebraska, despite being registered in multiple locations under the International Registration Plan.  See Berg. v. Liberty Mut. Ins. Co., 319 F.Supp.2d 933, 938 (N.D. Iowa 2004). Thus, the proper question is not whether an ADS is licensed, but rather whether it is required to be registered, and if so, it will likely be licensed as well.

What about the location of a Personal Mobility Unit? A vehicle is generally “principally garaged” in the physical location where it is kept most of the time. A vehicle that is kept in a single location for four months was “principally garaged” in that location even though the owner did not intend to reside there permanently. Chalef v. Ryerson, 277 N.J. Super. 22, 28, 648 A.2d 1139, 1142 (App. Div. 1994). A vehicle was not “principally garaged” in Michigan because it was not kept in Michigan “most of the time.” Frasca v. United States, 702 F. Supp. 715, 718 (C.D. Ill. 1989). But a rental truck leased for a three-day period was “principally garaged” in Virginia, not Maryland where it was licensed and registered, because the leasee kept the truck at its Virginia facility during the lease period. Hall v. Travelers Cas. Ins. Co. of Am., No. 1:16-CV-00173 (GBL), 2016 WL 5420571, at *5 (E.D. Va. Sept. 27, 2016). If the manufacturer owns the autonomous vehicle and leases it to different individuals in different states, where will that vehicle be principally garaged, and will it be subject to a compulsory or financial responsibility law or other motor vehicle insurance law in that state?

Finally, “auto” excludes “mobile equipment.” “Mobile equipment” encompasses an array of land vehicles, from bulldozers to power cranes to air compressors to any vehicle “maintained primarily for purposes other than the transportation of persons or cargo.” By definition, any autonomous vehicle, including a Personal Mobility Unit, is used to transport people and cargo and therefore will not constitute “mobile equipment.” In addition, the Commercial Auto Coverage Form excludes from the definition of “mobile equipment” those “land vehicles that are subject to a compulsory or financial responsibility law or other motor vehicle insurance law where it is licensed or principally garaged.” Those land vehicles are “autos.” As discussed above, autonomous vehicles, including Level 52 vehicles, will likely satisfy this exception to the definition of “mobile equipment” and will be considered “autos.”

Does Teleoperation Affect Coverage?

As of April 2018, California mandates a remote operator for autonomous vehicles testing on public roads without drivers. Cal. Code Reg. § 227.38. Remote operators must: (1) have the appropriate driver’s license for the type of autonomous vehicle; (2) not be seated in the driver’s seat of the vehicle; (3) engage and monitor the autonomous vehicle; and (4) be able to communicate with occupants in the vehicle by a communication link. The remote operator may also have the ability to “perform the dynamic driving task for the vehicle or cause the vehicle to achieve a minimal risk condition.”  Cal. Code. Reg. § 227.02(n).

At least one company will offer that remote control to assist with the testing and development of autonomous vehicles. Phantom Auto states that it provides a remote driver seated in a control room who can operate the vehicle remotely (described as “teleoperation”). Phantom Auto describes its services as “a teleoperation-as-a-service safety solution for all autonomous vehicles that includes an API [application program interface] for real time assistance and guidance, an in-vehicle low latency communication device, and a remote operator service.”  Phantom Auto enables a remote human operator to operate an autonomous vehicle when it encounters a scenario that it cannot handle on its own, allowing for optimally safe testing and deployment of autonomous vehicles.

The ISO Commercial Auto Form covers accidents “resulting from the ownership, maintenance or use of a covered ‘auto.’” Will insurers contend that an error in teleoperation does not “result from” the “ownership, maintenance or use of a covered ‘auto’” but instead “results from” remote operator negligence? Will they argue that other insurance policies should respond instead of a commercial auto policy?

We don’t yet know the answers to these questions, but insureds testing autonomous vehicles can determine the proper allocation of liability for remote operator error and tailor their policy language to address any liability. If teleoperation survives beyond the testing phase, all insureds using this service will need to assess the parties’ risk allocation and coverage options.

What about Exclusions?

As with the commercial liability policy, ISO’s Commercial Auto Form excludes coverage for “Expected or Intended Injury,” which it defines as “‘Bodily injury’ or ‘property damage’ expected or intended from the standpoint of the ‘insured.’” If the insured cannot operate the vehicle, can any circumstance arise that could arguably trigger this exclusion? At least from an operational standpoint, an insured could not expect or intend injury if the insured cannot operate the Personal Mobility Unit. Will insurers seize on allegedly defective maintenance rather than negligent operation to trigger this exclusion? Will they remove the exclusion as no longer applicable in an autonomous vehicle world? Only time will tell, but insureds and insurers alike should consider the impact of this exclusion.

What about Premium?

In February 2015, NHTSA reported that 94% of auto accidents are caused by human error.    Pundits predict that the frequency and severity of accidents will dramatically decline as autonomous vehicles become more prevalent (despite a messy intervening period when human drivers and partial and fully autonomous vehicles share the road). As frequency and severity falls, so does risk, thus potentially impacting premium. The complexity of autonomous vehicles may impact premium in the other direction, as early model autonomous vehicles will likely be extremely expensive to repair and replace. Insurers are no doubt (or should be) assessing autonomous vehicle’s potential impact on auto premium.

Where Next? Products Liability Insurance? Cyber Insurance?

Even as auto accident frequency and severity declines with increased automation, and premium eventually drops as repair costs normalize, commercial auto insurers may need to turn to other lines of coverage, such as products liability and cyber coverage, for revenue streams, and insureds may need to turn to these same policies for coverage for autonomous vehicles. We will discuss those policies in subsequent posts with a focus on Level 52 vehicles.

Mobile App Terms and Conditions Decision Clarifies Best Practices in App Designs to Support Enforcement of Contract ProvisionsThe Second Circuit issued a decision of interest to every company that utilizes a mobile app to interact with its customers. In Meyer v. Kalanick, the court enforced the mandatory arbitration provision in the Uber app. The court considered the app from the perspective of a “reasonably prudent smartphone user” and discussed parameters supporting enforceability of contract terms for mobile apps. The Second Circuit enforced the arbitration provision because the Uber app gave the user reasonably conspicuous notice of the Terms of Service (which included the arbitration provision), and the user gave unambiguous (albeit not express) consent to arbitration in light of the objectively reasonable notice of the terms.

Due to the ubiquity of smartphones and smartphone apps, the Second Circuit analyzed the Uber app from the standpoint of a reasonably prudent smartphone user who would understand the use of hyperlinks. Uber’s “uncluttered” payment screen contained the warning that “By creating an Uber account, you agree to the TERMS OF SERVICE & PRIVACY POLICY.” The court explained that the “capitalized phrase is bright blue and underlined and contains a hyperlink to a third screen containing a button that, when clicked displays the current version of Uber’s Terms of Service and Privacy Policy. The text, including the hyperlinks to the Terms and Conditions and Privacy Policy, appears directly below the buttons for registration.” The entire screen, including the notice of the Terms of Service, was visible without scrolling. The court noted that the sentence is in small font, but “the dark print contrasts with the bright white background, and the hyperlinks are in blue and underlined.” The court appreciated the simplicity of the payment screen, which included only credit card fields, buttons to register for a user account or to connect pre-existing accounts to the Uber account, and the warning with the hyperlink.

The court explained that a reasonably prudent smartphone user would understand that text that is highlighted in blue and underlined is hyperlinked to another webpage with additional information, and found the screen design and text reasonably conspicuous, thus giving the user constructive notice of its terms. The court also described the payment screen and Terms of Service as “temporally coupled” because Uber provides the Terms of Service during enrollment. The court concluded that “a reasonably prudent smartphone user would understand that the terms were connected to the creation of a user account.”

The court distinguished the Uber screen from the Amazon screen in Nicosia v. Amazon.com, Inc., because the Nicosia screen contained much more information and several buttons, and the notice of terms and conditions was not adjacent to the consent button: “This presentation differs sharply from the screen we considered in Nicosia, which contained, among other things, summaries of the user’s purchase and delivery information, ‘between fifteen and twenty-five links,’ ‘text . . . in at least four font sizes and six colors,’ and several buttons and advertisements.  Nicosia, 834 F.3d at 236-37. Furthermore, the notice of the terms and conditions in Nicosia was ‘not directly adjacent’ to the button intended to manifest assent to the terms, unlike the text and button at issue here.  Id. at 236.”

The Uber app decision provides useful guideposts for designing user interfaces for smartphone apps that include contractual terms, such as arbitration clauses: (1) implement a simple design with minimal text and few buttons; (2) ensure the visibility of the entire screen, including the hyperlink to the contract terms, without scrolling; (3) expressly warn that by creating an account, the user is agreeing to be bound by the linked terms; and (4) require agreement to the contract terms during enrollment (ideally before completing enrollment, but not later than simultaneously with enrollment).  Although the Uber app did not do so, smartphone apps can also require the user to scroll through the governing terms and conditions before accepting them to further support enforceability of those terms and conditions.

self driving carIf you weren’t already convinced, this Drive.ai video of a self-driving system demonstrating an extended drive, at night, in rainy conditions, without human intervention, shows how close we may be as a society in which human driving is a luxury activity.

Drive.ai is far from the only company developing and testing AV systems. Waymo, Alphabet’s AV subsidiary, logged more than 635,000 autonomous miles driven on California public roads in 2016 (all other competitors combined logged 20,286 autonomous miles in California). Recently, Ford announced plans to develop a Society of Engineers (SAE) Level 4 AV (no steering wheel, accelerator, or brake pedal) for use in a commercial application (such as ride-hailing of package delivery) by 2021. General Motors, which, along with Ford, strongly supported legislation to allow AV testing and deployment in Michigan beginning this March announced that it will begin production of its AV line early this year. Tesla’s Autopilot option, an advanced driver-assist feature that Tesla describes as offering “full self‑driving capability” but requires driver monitoring has logged over 200 million miles since it was released in 2015.  Tesla estimates that fully autonomous driving could be technically feasible in two years, even if the regulatory framework is not developed enough to allow fully AV yet.

The Regulatory Hurdle

With major industry players fully engaged, Tesla predicts that AV will be technically feasible for a wide-scale deployment in two years.  The adoption of AV regulations in all fifty states may take several years, but, given the potential increases in safety (an estimated 90% of accidents are caused by human error) and the substantial backing of the auto industry, implementing the regulatory framework is likely only a matter of time.  While the National Highway Traffic Safety Administration’s (NHTSA) AV Policy only provides voluntary guidelines for AV manufacturers and a model policy for state regulation of driverless technology, NHSTA affirmed that it intends to pursue an “ambitious approach to accelerate the highly automated vehicle revolution.”

Common Industry Terminology

NHTSA’s AV Policy adopted AV terminology from SAE, which categorizes vehicles by levels of human control.  Level 0 vehicles require constant human control, while Level 5 vehicles will need no help from humans. Level 3 vehicles utilize “conditional automation” in which “the human driver will respond appropriately to a request to intervene.”

Policyholder Impact

AV’s impact on the auto insurance industry – and by extension, commercial auto insurance – is not fully known. While most accidents are caused by human error, the rise of AV introduces a new source of error: misjudgments by the underlying AV system. Experts cannot yet draw any statistically significant conclusions about the safety of AV due to the relatively small sample size (experts estimate that it will take billions of vehicle miles to draw accurate conclusions). Uber is aggressively pursuing AV, but faced criticism when one of its AVs ran a traffic light in San Francisco. Uber blamed the incident on the human driver, who failed to intervene, rather than on the AV system itself. A NHTSA report following a fatal crash involving a Tesla with the Autopilot feature engaged, revealed that after Tesla introduced the Autopilot feature, the number of crashes dropped by 40%.

Assuming that Tesla’s Autopilot data reflects a true decrease in automobile accidents for AV, we might expect insurance premiums for all cars to drop. However, some experts contend that Level 3 AV (similar to Tesla’s Autopilot feature) may present more safety concerns than complete human control or Level 5 AV. Nidhi Kalra, co-director of the Rand Center for Decision Making Under Uncertainty, testified during a recent U.S. congressional hearing that “[t]here’s evidence to suggest that Level 3 may show an increase in traffic crashes . . . I don’t think there’s enough evidence to suggest that it should be prohibited at this time, but it does pose safety concerns.”  Experts question whether inattentive humans may fail to intervene quickly enough, causing the AV to take no action in a dangerous situation because it expects intervention, or the inattentive human may intervene and make the situation worse because of a lack of situational awareness. If this holds true, we might expect premiums to come down only for Level 5, and not for Level 3 or 4 AV.

In addition to the premium question, AV will likely also impact current coverage forms. In the United Kingdom, the Centre for Connected and Autonomous Vehicles recommended new regulations requiring auto insurance policies to include a separate coverage grant for AV. While the UK motor-vehicle insurance scheme is driver-specific rather than vehicle-specific, the authorized driver provision in U.S. policies poses similar questions for insureds. Commercial Auto Policy standard ISO form language for the “Who Is An Insured” provision defines “‘insureds’: [as] a. You for any covered ‘auto’. b. Anyone else while using with your permission a covered ‘auto’ you own, hire, or borrow except […] c. Anyone liable for the conduct of an ‘insured’ describe above but only to extent of that liability.” The standard ISO form defines “insured” as “any person or organization qualifying as an insured in the Who Is An Insured provision of the applicable coverage.” Insurers must determine whether AV falls within this definition of an “insured” and whether to revise or endorse the standard ISO form to capture AV exposure.  In addition, insurers must assess the impact of AV on the omnibus provision c., which defines as an “insured” “[a]nyone liable for the conduct of an “insured” described above but only to the extent of that liability.” This provision extends liability coverage to anyone vicariously liable for the actions of an insured; its impact on AV must be determined.

The Future of AV and Commercial Auto Coverage

For the foreseeable future, commercial auto policies will remain critical components of every commercial insured’s policy portfolio. Accidents will continue to occur, and insureds will need legal representation in any ensuing litigation, which implicates auto insurers’ duty to defend. Insureds will also need comprehensive and collision insurance to repair damaged vehicles.  Many insureds may drive a variety of vehicles, including Level 0 vehicles fully controlled by the driver, and will need traditional permissive-driver liability protection.

Numerous liability questions may arise in accidents involving Level 4 and 5 AV vehicles, including:

  • Who will be responsible for any liability? The passenger sitting in the driver’s seat but not driving? The product manufacturer? A third-party purportedly causing or contributing to the accident?
  • Will insurers use vehicle ratings for property damage coverage but use different metrics for liability coverage?
  • Will insureds see less frequency of claims on their business auto liability coverage, resulting in lower premiums and impacting insurance company premium volume?
  • And, of course, will standard and manuscript insurance policies change in response to this new world?