Workplace Robot Liability: 5 Critical Risks for Business

This week, the company that powers most of the world’s AI turned its attention to a quieter problem: keeping robots from hurting the people standing next to them. As Dr. Alex Wissner-Gross reported in the June 22 edition of The Innermost Loop, NVIDIA unveiled Halos for Robotics, billed as the first full-stack safety system for “physical AI” — robots that move, lift, and make decisions on a factory or warehouse floor. Agility Robotics is the first partner, bringing the system to facilities run by Amazon, GXO, and Toyota. It is a milestone for automation. It is also the moment workplace robot liability stops being a science-fiction question and becomes a line item on your risk register.

What NVIDIA’s Robot Safety System Means for Business

For most of the last decade, “the robots are coming” was a slogan, not a schedule. That changed fast. Humanoid and collaborative robots — “cobots” that share space with human workers rather than sitting behind a cage — are now being deployed at scale in warehouses, plants, and logistics yards. NVIDIA’s announcement matters because it tries to standardize how those machines stay safe, drawing on years of self-driving-car safety engineering and routing it toward robots that work an arm’s length from your staff.

Here is the part business owners should sit up for. NVIDIA built an inspection lab to help robot makers certify their systems against rigorous functional-safety standards such as IEC 61508 and ISO 13849 before third-party certification. In plain English, the industry is building a yardstick for what a “reasonably safe” workplace robot looks like. Once that yardstick exists, every business that deploys a robot will be measured against it — by regulators, by insurers, and, when something goes wrong, by a plaintiff’s lawyer.

The deployments are already here, not theoretical. The same week, JD.com’s founder said the company’s 700,000 couriers will be replaced by robots “sooner or later,” and General Motors began adding cobots to a Detroit truck plant. When robots move from novelty to infrastructure, workplace robot liability moves with them.

The Legal Impact: 5 Workplace Robot Liability Risks

There is no single “robot law.” Instead, a robot on your floor sits at the intersection of product liability, workplace-safety regulation, contract law, insurance, and labor law — and most businesses have never mapped that intersection. Below are the five places workplace robot liability lands first, and why each one deserves attention before the machine is bolted down.

1. The standard of care arrives before the regulations do

Lawmakers are slow; industry standards are not. When a respected vendor framework like NVIDIA’s Halos defines what a safe physical-AI system should include, courts treat that emerging consensus as evidence of the standard of care — the level of caution a reasonable business owes. A company that deploys a robot falling short of the recognized standard hands a negligence plaintiff a ready-made argument. Conversely, certifying to it becomes a defense. The exposure is real long before any statute is written, because juries decide what was “reasonable,” and the industry is now telling them.

2. OSHA’s General Duty Clause already covers your robots

You do not get to wait for a robot-specific rule. OSHA has no dedicated robotics standard, but it enforces the General Duty Clause — Section 5(a)(1) of the OSH Act — which requires a workplace “free from recognized hazards.” A robot operating near employees is a recognized hazard, and OSHA’s own Technical Manual points employers to consensus standards like ANSI/RIA R15.06 (and the collaborative-robot guidance in TR R15.606) as the compliance baseline. Ignore those standards and an OSHA citation can follow even without a rule that names robots — and that citation becomes evidence in any injury suit that follows.

3. Your vendor and integrator contracts probably don’t allocate the risk

A deployed robot usually involves at least three companies: the manufacturer, the systems integrator who installs and programs it, and the AI or software vendor whose model makes the decisions. When that machine injures someone, the fight over who pays is governed by the contracts you signed. Most robotics purchase and integration agreements were drafted from the seller’s side and are silent — or buyer-unfavorable — on indemnification, warranty, limitation of liability, additional-insured status, and who must carry which insurance. Reviewing those business contracts before signing is the cheapest risk reduction available.

4. Workers’ comp may not be the end of your exposure

Many owners assume workers’ compensation closes the book on a workplace injury. It often does not. Comp generally bars an employee from suing the employer, but it does not stop that employee from suing the robot’s manufacturer or integrator in product liability — and those defendants can then pursue the employer through “third-party-over” claims and contractual indemnity. The result is that a single robot incident can drag your business into litigation it thought it was immune from. Serious bodily-injury claims of this kind are litigation matters, handled by our colleagues at Howard Law Group.

5. Automation triggers employment and labor obligations

Replacing people with machines is not a purely operational decision — it is a legal one. Large-scale workforce reductions can trigger notice obligations under the federal WARN Act and state equivalents. In a unionized shop, automation can be a mandatory subject of bargaining, as GM’s cobot rollout and the UAW’s reaction this month illustrate. Getting ahead of severance, retraining commitments, and employment law obligations avoids turning an efficiency win into a labor dispute.

Underneath all five risks runs a quiet insurance problem: general-liability and workers’-comp policies written before autonomous machines were common may exclude or under-cover bodily injury caused by an AI system making its own decisions. Confirming coverage — and adding robotics-specific endorsements where needed — belongs on the same checklist.

What Howard East Clients Should Do Now

You do not need to slow your automation roadmap. You need to put a few guardrails up before the robot goes live. Three moves are worth making this quarter.

First, demand certification and document it. Before deployment, get written confirmation that the robot and its integration meet recognized safety standards (ANSI/RIA R15.06, ISO 13849, and the cobot guidance), and keep the records. That paper trail is your first line of defense if an injury is ever litigated.

Second, renegotiate the contract before you sign it. Have counsel review the purchase, integration, and software agreements for indemnification, warranty, limitation-of-liability, additional-insured, and insurance provisions. The time to allocate workplace robot liability is at the table, not in a courtroom.

Third, align your insurance and your workforce plan. Confirm your carrier will cover autonomous-machine injuries, and if your deployment displaces workers, map your WARN Act and bargaining obligations early. Businesses in regulated and physical-operations fields — from manufacturing and logistics to cannabis cultivation and processing operators automating their grow and extraction lines — face these questions the moment a robot touches the floor, and the same discipline applies whether you are planning a plant retrofit or a facility buildout. For a related look at how fast machine autonomy is reshaping exposure, see our analysis of autonomous AI liability and your business contracts, and review your corporate risk posture accordingly.

Frequently Asked Questions

Who is liable when a workplace robot injures an employee?

It depends. Workers’ compensation usually covers the employee’s claim against the employer, but the manufacturer, integrator, or AI vendor can still be sued in product liability — and they may pursue the employer in turn. Workplace robot liability is ultimately allocated by the procurement contract, the insurance policies, and the safety record.

Does OSHA regulate robots that work near people?

Yes. OSHA has no robot-specific standard, but it enforces the General Duty Clause and treats consensus standards like ANSI/RIA R15.06 as recognized. An employer that ignores them can be cited even without a dedicated rule.

Should we review our contracts before deploying robots?

Yes. Most robotics purchase and integration agreements are silent on who bears the risk when an autonomous machine causes injury. Indemnification, warranty, insurance, and certification terms should all be reviewed before signing.

This article is for informational purposes only and does not constitute legal advice, and reading it does not create an attorney-client relationship. Workplace robot liability questions are fact-specific and evolving; consult qualified counsel about your situation. Attorney Advertising.

Source: Dr. Alex Wissner-Gross, The Innermost Loop, June 22, 2026.