Nobody Was Driving.
Someone Is Still Responsible.

For a century, the first question after a truck crash has been the same: who was driving? In a growing stretch of the country, that question no longer has an answer. Class 8 tractor-trailers are now running fully driverless on public highways — the cab is empty, and a computer is doing the driving. That is not a prediction about the future. It is happening right now, mostly in Texas, and the companies behind it plan to scale fast. It raises a question that catches most families completely off guard: if there was no driver, who is responsible when the truck destroys a life?

The short answer is that the responsible parties do not disappear when the driver does — they move up the chain, from a person behind the wheel to the companies that built, sold, and deployed the machine. This is a plain-English walk-through of who those parties are, what the law actually says about a truck that drives itself, how the rules are shifting in 2026, and why the strongest evidence in these cases is the truck’s own memory.

Are driverless trucks really on the road already?

Yes — and more than most people realize. Aurora Innovation launched commercial driverless trucking in Texas in April 2025, running its first loads between Dallas and Houston with no human in the cab. By January 2026 the company reported more than 250,000 driverless miles with, by its own account, zero collisions attributed to its automated driving system. By mid-2026 it had tripled its network to roughly ten driverless routes — adding Fort Worth to El Paso, El Paso to Phoenix, and Dallas to Laredo — and said it expects more than 200 driverless trucks in operation by the end of 2026, running on Volvo and International Class 8 platforms. A second company, Kodiak Robotics, is running autonomous freight with Roehl Transport on the Dallas–Houston corridor and is moving toward its own driverless launch.

In the language of vehicle automation, these are SAE Level 4 systems — the vehicle handles the entire driving task within a defined area, with no expectation that a human will take over. That is a different category from the driver-assist and “self-driving” features on consumer cars, which still assume a person is supervising. We covered that distinction, and why it matters, in our explainer on self-driving car crashes. The leap with freight is one of physics: a fully loaded tractor-trailer can weigh up to 80,000 pounds, roughly twenty times a typical passenger car. When a machine that heavy gets a decision wrong, the result is rarely a fender-bender. It is the kind of catastrophic outcome — severe brain and spinal injury, amputation, death — that this firm handles.

If no one is driving, who is responsible for a crash?

Removing the driver does not remove the defendants. It changes who they are and what kind of claim you bring. Instead of a single claim about one person’s momentary negligence — a glance at a phone, a missed brake — a driverless-truck case looks up the chain to the companies that made the machine and its decisions. In most cases that means several potential defendants:

The automated-driving-system developer. The company that designed, trained, and validated the self-driving software — Aurora, Kodiak, or another — is the party whose product actually made the driving decisions. A claim that the system failed to perceive a hazard, misclassified an object, or chose an unsafe maneuver is a product-liability claim: the theory is that the automated driving system was defective or unreasonably dangerous. The truck manufacturer. The Volvo or International tractor the system was installed on is itself a product; a defect in the vehicle or in how the autonomy hardware was integrated can bring the manufacturer in. The motor carrier or authorization holder. The company that owns the fleet and put the driverless truck on that road is responsible for its deployment, maintenance, and the decision to run it in given conditions. Maintenance and monitoring contractors. Driverless fleets rely on sensor calibration, servicing, and remote operations centers; a failure in any of those functions can create its own line of negligence.

This is the same instinct that drives every serious truck case — identify every party in the chain rather than stopping at the obvious one. It is the same move that lets a driver-error case grow into a claim against the broker that chose the carrier or into a corporate case built on safety-management failures. With a driverless truck, the chain simply starts one link further back: not with who was driving, but with who built the thing that was.

“The software is the operator” — what Texas law actually says

Because most driverless freight today runs in Texas, Texas law is the current proving ground. Under Texas Transportation Code, Subchapter J (Sections 545.451 through 545.456), when an automated driving system is engaged, the system itself is treated as the “operator” of the vehicle for purposes of traffic and motor-vehicle laws, and the owner — or the state authorization holder — is the party issued any traffic citation. On first read that sounds like it might let a trucking company shrug and say, “the software was driving, not us.”

It does not. That statutory label decides who gets the ticket, not who pays for a catastrophic injury. Civil liability for a crash is a separate question, governed by ordinary negligence and product-liability law, and it can reach the system developer, the vehicle manufacturer, and the company operating the truck all at once. If anything, a rule that formally names the software as the operator cuts in the injured person’s favor: it aims responsibility squarely at the companies that designed and deployed that software. A human driver’s negligence is usually capped by that one person’s insurance; a defective-product claim against a technology company and a truck manufacturer is a different kind of case, backed by different resources.

The ground is shifting in 2026 — and it matters to your case

Two regulatory changes in 2026 are reshaping how these cases will be built. First, Texas tightened its own oversight. A 2025 state law, Senate Bill 2807 (effective September 1, 2025, with enforcement phasing in through 2026), now requires a company to obtain authorization from the Texas Department of Motor Vehicles before operating a driverless vehicle commercially. To get it, the operator must certify that it has given the Department of Public Safety a plan for how first responders should interact with the driverless vehicle, must register and insure the vehicle, must equip it with a recording device, and must show the vehicle can reach a safe condition if the automated system fails. Each of those requirements is also a checklist a lawyer can hold a company against after a crash: did it actually do what it certified it would do?

Second, the federal government moved the other direction — toward deregulation. In 2026 the National Highway Traffic Safety Administration advanced its Automated Vehicle Framework, including rulemaking to update Federal Motor Vehicle Safety Standards for vehicles built with no human controls (for example, revisiting the manual brake-pedal requirement) and to speed up exemptions for commercial automated fleets. The Federal Motor Carrier Safety Administration is separately updating its rules for automated-driving-system-equipped trucks, including how human-specific requirements such as hours-of-service apply when there is no driver to rest. The takeaway for an injured family is blunt: the rules are being loosened to let this technology scale quickly, which puts more of the accountability weight on the civil-justice system — on the ability to prove, case by case, that a specific driverless truck was operated or built unsafely.

The evidence is different — and there is more of it

Here is the part that surprises people: a crash with no driver often leaves more evidence than a crash with one, not less. A driverless truck is a rolling data recorder. To drive itself, it must continuously sense and log the world around it — the feeds from its cameras, radar, and lidar, the system’s moment-by-moment perception of the road and the objects on it, the decisions the software made, and the exact instant of any disengagement or fault. Texas requires authorized automated vehicles to carry a recording device precisely so this record exists. Where a human witness offers a shaky, secondhand account, the machine holds a detailed, timestamped account of what it saw and what it did in the seconds before impact.

That data is powerful, and it is also perishable. It sits in the control of the very companies that would be defendants, and like all crash data it can be overwritten or lost if no one moves to lock it down. The reasons the first 24 hours after a crash matter are magnified here, and the discipline is the same one that applies to a conventional truck’s event data recorder — except the volume and importance of the data are far greater. A preservation demand aimed at the autonomy developer, the carrier, and the truck manufacturer, sent immediately, is often the single most important early step in a driverless-truck case.

How the medical-legal team approaches a driverless-truck case

A case like this is proven on two tracks at once. On the liability side, when Alex Alvarez, Board Certified Civil Trial Lawyer, works a driverless-truck matter, the questions are engineering questions dressed as legal ones: What did the system perceive, and when? Was the object the truck hit something the software should have detected and classified? Did the vehicle attempt to stop or steer, and was that response adequate for its speed and weight? Was the truck authorized and maintained as the operator certified? Those answers live in the log data and in the company’s own safety documentation, and getting to them requires treating the case as the product-liability and corporate-negligence case it is.

The medical side runs in parallel. Herb Borroto, M.D., J.D., reads the injuries the way a physician does — and in a heavy-truck crash the injuries are the predictable physics of mass and speed: a traumatic brain injury, a spinal cord injury, an amputation, or a fatal outcome. Reading the medicine against the truck’s own record of what it did is what connects a line of software code to a human being’s permanent loss. That pairing — a physician who reads the imaging and a trial lawyer who builds the liability case — is how this firm approaches a crash where the defendant is, in part, a machine.

What this means for an injured person or family

If you or someone you love was catastrophically hurt or killed in a crash involving a commercial truck — and especially one that may have been operating autonomously — a few things follow. First, the fact that no one was driving does not mean no one is accountable; it usually means the responsible parties are companies, not a single driver, and identifying all of them is part of a complete truck-crash investigation. Second, the case is unusually time-sensitive, because the most important evidence is digital, perishable, and held by the potential defendants — the work of preserving it has to start immediately. Third, this is a genuinely new and technical area of law that most general practitioners have never touched; it calls for a firm comfortable with product-liability litigation and federal motor-carrier issues, not just routine insurance claims. The technology is moving faster than the public conversation about who answers when it fails. Making sure an injured family is not the one left holding that failure is the work.