There is no driver, so there is no driver to blame. That single fact moves your case out of ordinary negligence and into product liability.
You are probably suing the fleet operator and the manufacturer, and the case turns on data you do not have and they do.
It changes who the defendants are, what evidence matters, and how long you have before that evidence is gone.
Why “who was driving” is the wrong question
An ordinary car crash case asks whether a person behaved unreasonably. Did they speed, look away, misjudge a gap, run the light.
A driverless vehicle has no such person. The decisions were made by a perception and planning system, written months or years earlier, validated against test scenarios chosen by engineers, and deployed by a company that decided it was ready for public roads.
So the question is no longer “was the driver careless.” It becomes:
- Was the system defectively designed?
- Did it perform as designed, and was that performance itself the defect?
- Did the operator deploy it in conditions it was not validated for?
- Were the warnings and limitations adequately disclosed?
Those are product liability questions. They are answered with engineering evidence, internal validation records, and testimony from people who built the thing. That is a fundamentally different case from a two-car collision, and most firms are not set up to try it.
Who the defendants actually are
| Potential defendant | The theory | The obstacle |
|---|---|---|
| Fleet operator | Negligent deployment, inadequate remote monitoring, operating in unvalidated conditions, failure to remove a known-defective vehicle from service | They control the data and will resist producing it |
| Vehicle manufacturer | Design defect, manufacturing defect, failure to warn | Compliance with federal standards is argued as a shield |
| Autonomy software developer | Defective design of the perception or planning system | May be a separate corporate entity, sometimes offshore, and software has an unsettled relationship with product liability doctrine |
| Sensor or component supplier | A specific part failed | Component-level causation is hard to isolate |
| Remote operator or monitoring vendor | Negligent supervision or intervention failure | Their role and duty may be contractually obscured |
| Municipality or state | Roadway design, signal timing, or permitting | Governmental immunity and short notice deadlines |
| Another human driver | Ordinary negligence, in a mixed-fault crash | Often the smallest policy in the case |
Note what that table implies. This is a multi-defendant case with corporate parties, layered coverage, and cross-claims, which is exactly the structure of product liability litigation rather than a standard Houston car accident claim.
Is software a “product”?
This is the genuinely unresolved question underneath everything else, and it deserves a straight answer: the law has not settled it.
Product liability doctrine developed around physical things. Courts have long treated services and pure information differently from manufactured goods. Autonomous driving software is neither cleanly.
When the software is embedded in a vehicle at manufacture, treating the whole vehicle as the product is straightforward. When the behavior that caused the crash arrived in an over-the-air update pushed last Tuesday, it is much less clear. Is that a product defect, a negligent service, or something the law does not have a category for yet?
No court has given a durable answer. Anyone telling you otherwise is guessing.
The evidence problem, which is the real problem
An autonomous vehicle records far more than a conventional car. Camera, lidar, and radar streams. Perception outputs showing what the system classified your body as, and when, which is decisive in Houston pedestrian accident cases. Planning decisions. Disengagement and intervention logs. Remote operator communications. Software version at the moment of impact.
That data would resolve most of these cases in days.
You do not have it. The company does. And retention is governed by their internal policy, not by a rule written with injured people in mind.
Which means the single most important thing that happens in an autonomous vehicle case happens in the first week: a preservation demand specific enough to name the actual data types, sent before anything cycles.
A generic “please preserve all evidence” letter is not sufficient here. It has to name the sensor logs, the perception output, the planning stack telemetry, the software build, and the remote assistance records.
What Texas law does and does not say
Texas amended its Transportation Code to allow automated motor vehicles to operate without a human operator, and it assigns responsibility for compliance with traffic laws to the vehicle’s owner rather than to any occupant.
That is helpful for identifying a responsible party. It does not create a cause of action, define a standard of care for autonomous systems, or set out any evidence retention requirement.
There is no state or federal crashworthiness standard governing autonomous driving performance the way there is for brakes or airbags. There is no required data retention period. There is no mandated public reporting comparable to what aviation requires.
So the standard of care in your case will be established by experts, not by a statute. That is expensive, and it is why these cases require a firm willing to fund them properly.
What this means for you
If a driverless vehicle hit you, four things matter in the first 72 hours.
Photograph the vehicle, including every identifier. Fleet number, license plate, sensor housings, any visible markings. You need to establish which specific vehicle it was.
Get the incident number and every company contact. Operators dispatch representatives to scenes quickly. Write down names and titles. Do not give a recorded statement.
Preserve independent evidence. Nearby business and traffic camera footage, other drivers’ dashcams, witness contact information. The company’s own data is not the only record, and independent footage is leverage.
Get counsel moving on preservation immediately. This is the step that decides the case. Not the deadline. The data retention window.
Robert Law Group is building a practice in emerging technology torts, autonomous vehicles among them. The combination that matters here is complex product litigation experience plus the technical fluency to know what to ask for. Aimée Robert holds a degree in management of information systems from her studies at the United States Air Force Academy, then spent two decades litigating product and pharmaceutical cases against corporate defendants.
When the defense argument is “the system performed as designed,” someone has to be able to ask what it was designed to do, and who decided that was good enough.
Call 832-509-2303.
Future Lawsuits. The law before the law catches up.