Product liability suits against robotics manufacturers are hardly new.  Since a Michigan jury awarded $10 million for a worker killed by a malfunctioning robotic system in an automobile factory in the early 1980s, plaintiffs’ attorneys have invoked state product liability laws, which often impose strict liability and allow punitive damages, against makers of robotic systems.  Surgical, commercial, and agricultural systems have all drawn such suits.

What is new is how those laws apply as robotic systems are integrated with sophisticated software and artificial intelligence to create “robotics-as-a-service” offerings.  As robots operate closer to people with minimal supervision, product liability has become a leading risk and a frequent diligence topic for investors.

Courts Increasingly Treat Service-Related Robotics Systems as Subject to Product Liability Laws, with Potentially Expensive Consequences.

For decades, the law treated manufacturers and service providers differently.  A plaintiff injured by a product did not need to prove fault: the manufacturer and its suppliers could be strictly liable for a defect or inadequate warnings.  Service providers, by contrast, were judged under negligence principles.

That distinction is no longer so clear.  As the industry moves toward a “robotics-as-a-service” model—in which a system is placed at a customer site and supported by remote monitoring, software updates, and maintenance)—courts have increasingly been willing to apply product liability principles to software and AI-driven systems.

The consequences can be significant.  In August 2025, a Miami jury returned a $329 million verdict, including $200 million in punitive damages, after Tesla’s Autopilot system allegedly malfunctioned in a fatal crash.  In March 2026, a Los Angeles jury found Meta liable for $6 million due to the design of its algorithm.  And in June 2026, Waymo recalled nearly 3,900 robotaxis after a software glitch caused vehicles to disregard road closure signs.

Investors and Regulators Respond: Ignore Product Liability Laws at Your Own Risk.

Investors have taken notice of this legal risk.  Product liability exposure now appears routinely in robotics diligence, with attention to insurance adequacy, indemnity terms, incident history, regulatory posture, and safety documentation.

Regulators also have taken note, particularly in the European Union. The EU AI Act entered into force August 1, 2024, with main high-risk requirements effective August 2, 2026, though June 2026 amendments delayed certain obligations to December 2027 or August 2028.  The revised Product Liability Directive, which expressly brings software and AI within strict liability, took effect December 8, 2024 and must be transposed by December 9, 2026.  The Waymo recall, meanwhile, was filed with the National Highway Traffic Safety Administration.

So What’s a Robotics Company to Do?

Each state’s product liability law has its own nuances, but in general those laws allow plaintiffs to bring:

• Failure to Warn Claims: These allege that the manufacturer knew or should have known of a risk and failed to communicate it adequately.  They require no proof of flawed engineering; instead a plaintiff must show only that the company failed to tell users of a potential risk or incapability of a product.  Operational design domains (the conditions under which a system is validated) are hard to state plainly, and warnings must reach purchasers, safety managers, supervisors, and bystanders.
• Manufacturing Defect Claims: These allege that a particular unit departed from its intended design, such as a miswired sensor, an out-of-tolerance gearbox, or an off-specification battery cell.  In robotics, a small physical deviation can produce a large behavioral deviation and significant liability.
• Design Defect Claims: A plaintiff may also allege that a correctly built product is inherently defective by design. Such claims turn on whether the product performed as safely as an ordinary consumer would expect (the “consumer expectation” test) or whether the design’s risks outweigh its benefits in light of a feasible, safer alternative design (the “risk-utility” test).  Neither test maps cleanly onto autonomous systems, inviting arguments that a different sensor, motion planner, or speed limit would have avoided the accident.

Companies can begin protecting themselves against such claims with several proactive steps:

• Review marketing materials carefully, particularly claims that systems are “autonomous” or “safe around people,” which can support arguments of reliance beyond validated capabilities.
• Keep contemporaneous design records documenting safety trade-offs, standards applied, and hazard analyses, which are critical to defending design defect claims.
• Implement supplier qualification, inspection, process control, end-of-line functional testing, and traceability, and set retention policies for sensor, telemetry, and model-inference data before an incident occurs.
• Draft supply, integration, and deployment agreements with care, addressing indemnity scope, defense obligations, liability caps and carve-outs for personal injury, insurance requirements, recall cost allocation, and post-incident data access.
• Review insurance coverage regularly to confirm adequate protection against product liability claims.

Outcomes often turn on the quality of documentation, the discipline of release and incident processes, and the care with which risk is allocated by contract.  Companies that build these capabilities early gain a competitive advantage in deployment, diligence, and customer confidence.

Foley Hoag’s robotics practice advises startups, manufacturers, integrators, and investors on products liability issues, from safety program design and supply chain contracting through regulatory engagement and defense of claims.  Please contact the authors or your regular Foley Hoag attorney.