Product Liability for Robots and Autonomy Systems

Identify the product and defendant

Chapter 82 defines a products liability action broadly by the alleged harm and defective product, including claims pleaded under several theories. It separately defines seller and manufacturer. A manufacturer includes specified designers, assemblers, rebuilders, producers, and other participants who place a product or component into the stream of commerce. The label developer or integrator is not a substitute for that statutory analysis. [1, § 82.001]

A robot installation can contain several potentially relevant products and services. The arm, gripper, sensor, mobile base, safety controller, autonomy software, and integration work may have different suppliers. The factual investigation should identify which item or conduct allegedly produced the harm. A complaint about the overall fleet should not obscure whether the asserted failure arose in design, manufacture, installation, maintenance, operating instructions, or actual use.

Distinguish the defect theories

Texas product cases recognize manufacturing, design, and marketing-defect theories. A manufacturing theory concerns a departure from the intended product. A design theory challenges the chosen design. A marketing theory concerns warnings or instructions. The theory matters because the proof and applicable statutory provisions differ. Timpte explains the design-defect framework rather than treating all accidents as proof of a defect. [2]

For robotics, a misassembled brake, an allegedly inadequate obstacle-detection design, and an instruction that omits a known setup hazard present different questions. These examples identify possible theories, not findings that any such feature is defective. The evidence must connect the particular condition to the injury.

Design defects and safer alternatives

For a claim governed by Section 82.005, the claimant must prove a safer alternative design and that the defect was a producing cause of the injury, damage, or death. The statutory alternative must, in reasonable probability, prevent or significantly reduce the risk without substantially impairing utility. It must also have been economically and technologically feasible when the product left the relevant manufacturer or seller's control, using existing or reasonably achievable knowledge. [1]

That framework makes a specific proposal important. An assertion that a robot should simply be safer does not identify an alternative design or its feasibility. Possible changes such as a different protective system, braking arrangement, operating envelope, or sensor configuration require evidence about the relevant risk, utility, cost, and technical circumstances. Timpte also addresses the requirement that the design be unreasonably dangerous under the applicable risk-utility analysis. [2]

Section 82.005 contains exclusions, including a drug or device as defined in the federal Food, Drug, and Cosmetic Act. A surgical-robot case therefore should not receive the ordinary industrial-product statutory analysis without first addressing that exclusion and any governing device-specific law. The exclusion is not itself a ruling that the medical product cannot be defective. [1, § 82.005(d) and (e)]

Proving causation in an autonomous system

In Nissan Motor Co. v. Armstrong, the Texas Supreme Court rejected the proposition that unintended acceleration alone proved a vehicle defect. The opinion addresses competent evidence of the claimed defect and restrictions on using other incidents. It supplies a useful caution for autonomous machinery: an unexpected movement does not identify its own legal cause. [3]

An investigation may need the commanded action, sensor inputs, software and map versions, communications state, operating limits, maintenance history, human interventions, and physical condition. A log entry showing the controller requested movement answers a different question from why that request occurred. A missing detection can reflect several possible causes, which must be evaluated rather than assumed.

Evidence from other events can be relevant for particular purposes, but Armstrong requires attention to similarity and admissibility. A list of unrelated robot incidents does not automatically establish the alleged defect in the machine involved. [3]

Sellers, integrators, and indemnity

Section 82.003 generally protects a nonmanufacturing seller unless the claimant proves an enumerated basis for liability. Those bases include specified participation in design, modification, installation, control of inadequate warnings, incorrect factual representations, actual knowledge of a causative defect, and certain manufacturer insolvency or jurisdiction circumstances. The chapter contains additional qualifications, including its interaction with motor-vehicle dealer law. [1]

An integrator should therefore examine what it actually designed, changed, installed, or represented. A distributor that merely supplied equipment and an integrator that redesigned a protective arrangement may present different statutory facts.

Section 82.002 separately requires a manufacturer to indemnify an eligible seller for losses arising from a products liability action, except for losses caused by the seller's own specified wrongdoing or independently actionable conduct. The statute addresses notice and the scope of recoverable loss. Its indemnity is additional to duties established by contract or other law. A negotiated indemnity should be read alongside that statutory framework. [1]

Compliance is relevant but conditional

Section 82.008 creates a rebuttable presumption for specified formulation, labeling, or design claims when the defendant establishes compliance with mandatory federal safety standards that applied at manufacture and governed the risk that allegedly caused harm. It also addresses a separate federal premarket-approval circumstance and ways to rebut the presumptions. [1]

The presumption does not extend to manufacturing flaws or defects under subsection (d). A voluntary industry standard, a general statement of regulatory compliance, or a standard unrelated to the alleged risk is not automatically the statutory showing. Federal preemption is also a separate legal question; a Texas evidentiary presumption should not be described as complete immunity. [1]

Updates, operating conduct, and shared responsibility

A later software release can change the configuration relevant to the incident. The investigation should identify who supplied and authorized it, what it changed, when it reached the machine, and what testing or instructions accompanied it. Existing vehicle recall and consumer-product reporting statutes may create specific corrective-action duties. Those statutes do not establish one universal Texas duty to update every kind of robot. [5]

Operational conduct can matter alongside a product claim. Maintenance, installation, route selection, supervision, and modifications may support distinct allegations. Chapter 33 addresses proportionate responsibility where it applies; it does not supply the underlying duty or prove a defect. The allocation analysis begins after the relevant causes of action and responsible participants are identified. [4]

Hypothetical warehouse collision

A mobile robot strikes an employee after a new payload attachment is installed. The investigation should preserve the attachment design, installed software, sensor field, route, stop settings, maintenance records, and instructions. The evidence may support a product theory, integration theory, operating theory, or a combination. The collision alone does not establish which participant is liable. [1] [2] [3]