Georgia already permits fully driverless vehicles. As autonomous commercial trucking scales, liability will increasingly turn on corporate deployment decisions, product design, and the electronic record of what the system perceived and decided before a crash.

Driverless Class 8 tractor-trailers – the heaviest category of commercial trucks, generally vehicles with a gross vehicle weight rating above 33,000 pounds – are moving into commercial operation. Aurora Innovation began commercial driverless freight service in Texas in 2025 and, in July 2026, announced a new fleet of International LT-based driverless trucks operating without a person behind the wheel. Aurora reported that it was fully allocated to exit 2026 with more than 200 driverless trucks in operation. Kodiak AI reported 35 customer-owned driverless vehicles operating commercially in the Permian Basin at the end of the second quarter of 2026 and continues to target driverless long-haul operations by year-end. These are no longer merely test programs; autonomous heavy-truck deployment is beginning to scale.

The vehicles and freight operations are real, but the civil-liability framework remains unsettled. When a commercial vehicle operates without a human driver, the traditional question of who was driving gives way to more complex questions: what did the automated system perceive and decide, and which entities were responsible for the design, deployment, maintenance, and operation decisions connected to the failure?

That distinction matters in Georgia. Commercial trucking cases have always required counsel to look beyond the person behind the wheel and examine the motor carrier’s decisions, the condition of the equipment, the companies involved in the movement of the freight, and the evidence those entities control. Autonomous trucking does not eliminate that analysis. It expands it.

Georgia Already Permits Fully Driverless Vehicles

Georgia authorized fully autonomous vehicle operation in 2017 through Senate Bill 219, codified principally at O.C.G.A. § 40-8-11. The statute permits operation of a fully autonomous vehicle with the automated driving system engaged and without a human driver present, provided the statutory conditions are satisfied.

Among other requirements, the vehicle must be capable of operating in compliance with Georgia’s motor-vehicle and traffic laws; must have been certified by the manufacturer at the time of manufacture as compliant with applicable federal motor vehicle safety standards unless an exemption applies; must be capable of satisfying Georgia’s accident-reporting and scene obligations; must be able to achieve a minimal-risk condition if the automated driving system cannot perform the entire dynamic driving task within its operational design domain; must carry the required liability coverage or qualifying self-insurance; and must be properly registered as a fully autonomous vehicle or lawfully registered outside Georgia.

Two features are particularly important in future litigation. First, Georgia law does not require the person responsible for operating a fully autonomous vehicle to hold a driver’s license when the automated driving system is engaged. The traditional focal point of a motor-vehicle case – the licensed human operator – may therefore be absent entirely. Second, Georgia later amended the statute to preserve the application of specified state consumer-protection laws to autonomous vehicles.

What Georgia’s autonomous-vehicle statute does not do is allocate civil responsibility after a crash. It does not decide whether the motor carrier, autonomous-driving-system developer, truck manufacturer, component supplier, maintenance contractor, remote-support provider, or some combination of those entities bears the loss. Existing negligence, product-liability, apportionment, and evidence-preservation law will have to do much of that work.

The Federal Framework Is Still Developing

Federal regulation is developing around autonomous commercial vehicles, but it has not produced a comprehensive civil-liability regime. The Federal Motor Carrier Safety Administration (FMCSA) has stated that its regulations do not require a human driver or operator to be physically present when a commercial motor vehicle is equipped with a Level 4 or Level 5 automated driving system and, for Level 4, is operating within its operational design domain. FMCSA has also made clear that motor carriers operating ADS-equipped commercial vehicles remain subject to federal safety oversight.

The regulatory gaps are practical, not theoretical. Traditional rules assume that a human can perform tasks such as placing warning devices around a disabled truck. In 2026, FMCSA issued a limited waiver allowing Aurora – and other qualifying Level 4 motor carriers that satisfy the waiver’s conditions – to use specified cab-mounted warning beacons instead of conventional roadside warning devices. The waiver illustrates how existing safety rules are being adapted to driverless operations rather than replaced wholesale.

The durable point for civil litigation is that federal safety regulation and state tort law are developing on separate tracks. Federal rules may define operational obligations and supply evidence relevant to breach, but there is still no single federal statute that tells a Georgia court how to allocate civil liability when an autonomous commercial truck causes injury.

The Defendant List Expands

In a conventional trucking case, the core defendants are usually the driver and motor carrier, with brokers, shippers, maintenance companies, manufacturers, or others added when the facts support it. In an autonomous case, the list changes shape.

The motor carrier or fleet operator remains central. Automation does not erase the carrier’s operational decisions. A carrier still chooses where and when the truck will operate, whether conditions fall within the system’s operational design domain, how the equipment will be inspected and maintained, how software updates and safety notices will be handled, what remote-support structure will exist, and what performance data will trigger removal of a vehicle or system from service.

The automated driving system developer becomes a potentially central defendant, but the legal theory may depend on how the technology is supplied and integrated. Perception, prediction, planning, and control functions that once belonged to a human driver are performed by software and hardware. A failure to detect a pedestrian, classify a stopped vehicle, predict another road user’s movement, select a safe maneuver, or execute the commanded braking response may implicate the developer, a component manufacturer, the vehicle integrator, or more than one of them.

The truck manufacturer remains potentially responsible for conventional vehicle systems such as braking, steering, and structural performance, as well as integration failures where the autonomous-driving system and base vehicle are supplied by different entities. Sensor manufacturers, mapping and localization providers, connectivity vendors, and calibration contractors may also matter depending on the failure mode. If remote assistance (off-vehicle support to an automated truck) or teleoperation (remote human control of the vehicle) is involved, the entity providing that service – and in some circumstances the person performing it – may become part of the liability analysis.

That expanded corporate structure is one reason autonomous-truck litigation is unlikely to become simpler merely because the human driver disappears. The driver’s conduct may be replaced by a chain of design, deployment, maintenance, monitoring, and escalation decisions distributed among several companies.

Negligence and Product Liability Will Run in Parallel

Georgia provides both routes, and serious autonomous-vehicle cases are likely to plead them together.

On the product side, O.C.G.A. § 51-1-11 provides a strict-liability framework for injuries caused by new personal property that was not merchantable and reasonably suited to its intended use when sold. Georgia design-defect law applies the risk-utility analysis articulated in Banks v. ICI Americas, Inc., 264 Ga. 732, 450 S.E.2d 671 (1994), which weighs the risks inherent in a design against its utility and permits consideration of a feasible alternative design. Georgia’s product-liability statute of repose and its exceptions will also require careful analysis. Autonomous technology adds a threshold issue that conventional vehicle cases do not always present: whether the allegedly defective functionality is part of a product placed into the stream of commerce, a later software or service function, or an integrated combination of hardware and software. That characterization may affect which product-liability theories are available and against whom.

Where product-liability law applies, the framework maps naturally onto many autonomous-driving failures. If a system can activate in conditions it was not designed to handle, the scope and enforcement of its operational design domain – the conditions under which the system is designed to perform the driving task – may become a design question. If its perception system repeatedly fails to recognize a category of road user or roadway condition, the defect analysis may focus on architecture, training, validation, integration, or the availability of a safer alternative. Warning and representation theories may arise when a system’s capabilities are described in a way that encourages reliance beyond what the technology can safely support.

Negligence claims focus on different conduct: the carrier’s deployment decisions, maintenance and calibration practices, monitoring, response to safety data, training of remote personnel, and decisions about when to keep a vehicle in or remove it from service. Federal safety requirements remain relevant to those operational duties even when no human driver is sitting behind the wheel.

What Benavides v. Tesla Signals – and What It Does Not

The most instructive jury verdict to date comes from passenger-vehicle technology, not autonomous commercial trucking. In Benavides v. Tesla, a federal jury in Miami found Tesla 33 percent responsible for a 2019 crash involving Autopilot and a distracted human driver. The jury awarded $129 million in compensatory damages before allocation of fault and $200 million in punitive damages. The district court entered a final judgment of $242.57 million against Tesla and denied Tesla’s post-trial motions in February 2026. Tesla filed its appeal in the Eleventh Circuit in March 2026, where the case remains pending.

Benavides is not a Georgia autonomous-truck case, and it should not be treated as one. But it illustrates issues likely to recur: whether a driving system was allowed to operate where it should not have been used; what the manufacturer or developer represented about system capability; how fault is allocated when technology and human conduct overlap; and how electronic vehicle data can become central to proving what actually happened.

The case is especially relevant to evidence. The parties fought over vehicle data and the manner in which it was recovered and used. Autonomous commercial vehicles will generate vastly more information about the machine’s perception and decision-making than a conventional truck, making early preservation and technically competent discovery essential.

In an Autonomous Truck Case, the Data May Be the Case

A human driver can testify about what he saw, what he thought another vehicle would do, and why he braked or turned. An autonomous truck records versions of those questions in data.

Potentially critical evidence includes raw and processed camera, radar, and lidar (light detection and ranging) data; object-detection and classification outputs; prediction data showing what the system expected other road users to do; planning outputs showing the maneuver selected by the system; control commands and actuator responses; fault and disengagement logs; remote-assistance sessions and communications; the exact software and firmware build operating at the time; over-the-air update history; sensor calibration and maintenance records; and the simulation and validation testing relevant to the scenario at issue.

The evidence inquiry should not stop with the truck. At the developer and carrier level, counsel may need safety-case materials, operational-design-domain definitions and changes, prior similar events, known failure modes, internal escalation criteria, route-approval records, weather and roadway restrictions, remote-support protocols, and data showing how the system performed in comparable conditions.

In my commercial trucking practice at Goldstein Hayes Trial Lawyers, preservation disputes often begin with evidence controlled by motor carriers and other corporate defendants. Autonomous vehicles add new data layers and new custodians, but the litigation principle is familiar: identify the evidence before it disappears, identify who controls it, and preserve it with enough technical specificity that a request cannot be answered with only a conventional event-data-recorder download while the system-level evidence remains elsewhere.

Georgia’s spoliation law makes timing important. Under Phillips v. Harmon, 297 Ga. 386, 774 S.E.2d 596 (2015), the duty to preserve may arise when litigation is reasonably foreseeable to the party controlling the evidence, and notice can be constructive rather than formal. A catastrophic collision involving a driverless commercial vehicle can make preservation an immediate issue, particularly where the entities controlling relevant evidence begin their own investigation or data review.

The Practical Consequences for Georgia Catastrophic-Injury Cases

The first consequence is insurance. Coverage may exist through the motor carrier, fleet owner, technology developer, manufacturer, or excess programs, while contractual indemnity provisions redistribute the economic burden among defendants. Identifying every policy and contractual relationship may become as important as identifying every potentially responsible entity.

The second is apportionment. Georgia’s O.C.G.A. § 51-12-33 makes the identification of responsible persons and entities a substantive strategic issue. In an autonomous-truck case, a plaintiff may confront a carrier blaming the ADS developer, a developer blaming vehicle hardware or maintenance, a manufacturer blaming integration, and multiple defendants pointing to a remote operator or another road user. The technology does not eliminate fault allocation; it creates more places for fault to be assigned.

The third is expertise. These cases may require software and systems engineers, perception and sensor specialists, human-factors experts for remote supervision, and experts capable of evaluating validation and simulation evidence, in addition to conventional accident reconstruction, trucking-safety, medical, and economic testimony.

The fourth is speed. The truck will usually be recovered by an operator or corporate partner. Relevant data may reside on the vehicle, in remote servers, with the ADS developer, with the carrier, or across several vendors. Preservation decisions that once could be made after a routine investigation may need to be made within days.

The Underlying Question Has Not Changed

Automation removes the driver from the cab. It does not remove human and corporate decision-making.

Someone decided where the system could operate, under what weather and roadway conditions, at what speed, with what maintenance and calibration intervals, under what monitoring regime, and with what threshold for taking a truck or software build out of service. Someone decided when the technology was ready for public highways and how its capabilities would be described to customers, regulators, motor carriers, and the public.

Those decisions are made by identifiable entities, and increasingly they are documented in data. Autonomous trucking changes the evidence and expands the potential defendants, but it does not change the central task of catastrophic transportation litigation: reconstruct what happened, determine what the system knew, and follow the chain of decisions backward to the companies and people that designed, deployed, maintained, monitored, and ultimately put the vehicle on the road.