Last updated 4 months ago. Our resources are updated regularly but please keep in mind that links, programs, policies, and contact information do change.
Hyundai’s Savannah, Georgia factory will deploy 30,000 humanoid robots by 2028. These 330-pound machines will walk through assembly lines, lift 110 pounds, and work alongside human employees. The Atlas robot from Boston Dynamics is designed to work in spaces originally built for humans.
What federal workplace safety rules govern these robots? The answer: almost none that were written with humanoid robots in mind.
Existing Rules Were Written for Different Machines
OSHA’s robot safety standards were written for different machines. The basic rule is simple: keep robots and humans in separate areas, with safety cages maintaining that boundary.
That model works for a welding arm bolted to a factory floor. It doesn’t work for the Boston Dynamics machine.
The robot is designed to walk through dynamic manufacturing environments. It reaches into areas where human workers operate. The entire point is human-robot collaboration in shared workspaces—exactly what existing regulations don’t address.
A walking machine performing parts sequencing on an assembly line can’t be contained. It needs to move through aisles, work around obstacles, and operate near human coworkers. Physical barriers become impractical when the robot’s purpose is working in human-occupied spaces.
Hyundai’s Deployment Schedule
Hyundai plans to manufacture 30,000 units annually by 2028. The deployment schedule is specific: parts sequencing and material handling in 2028, component assembly by 2030, eventually expanding to heavy lifting and dangerous tasks. Hyundai frames this as “human-centered automation”—robots handle repetitive, high-risk work while humans move to supervisory roles.
The technical specs make clear why this differs from previous factory automation. The machine has capabilities designed for real factory conditions.
International Safety Standards Not Required in the U.S.
Other countries have moved faster on this. Boston Dynamics could adopt international specifications—but that would be a choice, not a requirement.
Hyundai emphasizes that robots will undergo “extensive safety validation” at the training center before deployment, with a focus on process-by-process validation and safety benefits, though the company has not explicitly detailed formal OSHA compliance testing or certification procedures or specific alignment with regulatory requirements. Specific details about how hard the robot can push, how close it can get, or adherence to international collaborative robot standards haven’t been disclosed.
Collision and Operational Hazards
The machine weighs 330 pounds and can lift 110 pounds with substantial force. A collision at even reduced speed could cause serious injury.
Traditional industrial robots operate behind barriers. If a worker gets injured, they’ve typically breached a safety barrier or disabled a safety interlock. Liability is clear—the worker violated protocols.
Factory robots create a different situation. Workers and robots share the same space by design. A collision doesn’t happen because someone breached a barrier. It happens because both are doing their jobs in proximity.
The machine can be controlled three ways: autonomously, through remote control by an operator, or via tablet interface. Each mode has distinct safety implications. Autonomous operation depends on reliable human detection and appropriate responses. Remote operation depends on communications quality and operator attention—delays in communication could cause accidents. Tablet control requires an operator to maintain awareness through limited sensor feeds.
Each unit carries substantial electrical power stored in its battery. Batteries overheating and catching fire in industrial equipment can create fire and explosion hazards. Workers near charging stations need appropriate training, but what training? No federal standards specify what that should include.
A traditional industrial robot, once disabled, stays relatively stationary. A disabled walking machine still presents hazards. Workers performing maintenance on hydraulic, electrical, or mechanical systems could face sudden movement or energy release.
If a machine malfunctions and starts moving unpredictably, workers need to know how to evacuate or disable it. The complexity of these procedures, combined with the rarity of walking robot deployments, means workers won’t have comparable experience to draw on.
Unclear Legal Liability
Say a walking machine seriously injures a worker at the Savannah plant. Who’s liable? Boston Dynamics as manufacturer? Hyundai as operator? The worker’s supervisor? The worker themselves for violating safety protocols—assuming any clear protocols exist?
Traditional workplace injury liability flows through workers’ compensation, which gives injured workers set benefits, but usually prevents them from suing their employer. When injuries involve products or equipment defects, workers can sue the manufacturer if the robot is defective.
Courts haven’t established clear rules for who’s responsible when robots cause injuries. The few significant autonomous system accidents that have been litigated—mostly involving Tesla’s autopilot—have produced mixed outcomes.
A walking machine operates in a controlled setting where the operator, manufacturer, and workers all have defined relationships. That bounded environment might simplify liability determination—or it might create heightened expectations about hazard prevention.
If the machine operates fully autonomously, responsibility rests primarily with the manufacturer and operator. If it’s remotely controlled, the remote operator bears significant responsibility. If it’s partly controlled by the robot and partly by a human operator, responsibility is distributed. Current injury laws may not cover these situations where figuring out what caused the injury is complicated by interactions between human operators, autonomous systems, and infrastructure.
Worker Training Gaps
Workers trained on traditional industrial robots have fundamentally different experience than what’s needed for walking machines. Traditional robot work emphasizes maintaining distance from restricted areas. Working with walking machines requires understanding robot behavior, recognizing hazard signs, responding to unexpected actions.
These machines have never been used in factories before. Hyundai’s training center began operations in 2026, with deployment not planned until 2028. Hyundai has limited ability to draw on proven training methodologies or employ trainers with extensive workplace experience with walking robots.
Workers assigned to these areas might lack appropriate preparation for machines with capabilities never before deployed commercially. An assembly line worker trained on traditional robots knows robots operate only in defined zones, move only when activated, respond to emergency stops. A worker alongside a walking machine must understand the robot can move through factory spaces, work around obstacles including humans, respond to sensor inputs rather than following pre-programmed paths, and potentially act in ways the worker didn’t anticipate.
Workers with limited automation experience or those transitioning from other roles face particular vulnerability. OSHA doesn’t specify credentials workers must have before working near walking robots, leaving qualification decisions entirely to operators. An operator eager to meet deployment timelines might assign workers with abbreviated training rather than thorough preparation.
Workers often have limited leverage to demand additional training or refuse assignments involving machines they don’t understand. In the absence of regulatory guidance, workers lack formal basis to demand training regulators haven’t identified as needed. Operators control training decisions while workers bear injury risks if training proves inadequate.
Worker Transitions and Employment Equity
Hyundai describes walking robot deployment as enabling workers to transition from repetitive physical labor to “higher-value supervisory and training roles.” But the implementation of such transitions remains uncertain.
Some workers might successfully transition to supervising, training, or supporting these machines. Others—particularly those with limited technical education or language skills in a plant with diverse workforce backgrounds—might struggle. They could accept lower-value assignments or face employment consequences.
The regulatory framework for ensuring equitable transitions is underdeveloped. OSHA focuses on workplace safety, not worker displacement, retraining, or employment equity. Companies implementing walking robots might voluntarily invest in retraining programs, but without regulatory requirements, retraining investment competes with other capital allocation decisions.
A company motivated to maximize return on investment might minimize retraining expenses. Management might choose to hire new workers with skills aligned to supervisory roles rather than investing in retraining existing workers.
In September 2025, an immigration enforcement operation at Hyundai’s Georgia battery plant resulted in detention of nearly 500 foreign workers. The timing of accelerated robot deployment following documented workforce vulnerabilities is notable.
Regulatory Timeline
Hyundai’s Robot Metaplant Application Center opened in 2026. First-phase deployment for parts sequencing is planned for 2028. Expansion to complex assembly tasks by 2030.
If OSHA intends to develop specific guidance for walking robots before commercial deployment at scale, the agency must initiate standards development in 2026 or early 2027 to complete guidance before 2028.
In January 2026, OSHA released proposals to remove or simplify safety rules, including proposals to remove certification requirements and simplify regulations, reflecting the administration’s emphasis on reducing regulatory burden. This environment seems unlikely to produce aggressive new safety standards for emerging technologies.
More likely: OSHA continues relying on existing general standards with voluntary adoption of international standards by forward-thinking companies. Workers and regulators learn from deployment experience. Liability litigation and workplace injuries drive subsequent regulatory development.
This pattern—emerging technologies proceeding to deployment with regulatory frameworks developing reactively—has characterized other automation deployments. It creates risk of preventable worker injuries.
Georgia has its own workplace safety agency instead of using federal OSHA. The Hyundai plant falls under state jurisdiction. If state regulators become concerned about walking robot safety before federal OSHA develops guidance, the state could potentially require specific safety measures. But unless workers or unions push for it, state regulators are unlikely to move ahead of federal standards.
Recommended Actions
Workers at plants deploying walking robots will perform tasks alongside machines with capabilities exceeding traditional industrial robots and behaviors workers may not understand or anticipate.
Training frameworks haven’t been established through regulatory processes. Training decisions rest with operators facing competing incentives about resource allocation. Legal liability for accidents remains unsettled, potentially shifting accident investigation from regulatory processes focused on worker protection to litigation focused on financial liability.
OSHA should quickly create safety rules for walking robots. These rules should cover how the robot must be built and how it must behave. Force limitations and sensor specifications need clear standards. Facility design and operational requirements for safe human-robot interactions must be established. Regulators should require companies to train workers and verify they understand the robots. Maintenance and emergency response procedures need specification. Companies should be required to track and report accidents and near-misses.
This guidance should adopt international safety standards for collaborative robots while adapting them to American regulatory frameworks. Clear safety standards should be set that companies must pass before deploying walking robots in production.
Boston Dynamics and Hyundai should establish safety frameworks exceeding minimum legal requirements. Both companies should recognize that novel technology creates risks regulatory standards haven’t addressed. Rigorous internal safety testing should be conducted and specific parameters for force limits, collision detection, and operator training should be established. Systems to report accidents and close calls during deployment should be created.
Operators should invest substantially in worker training. Workers trained on old robots need different training for walking machines. Training should address not only emergency procedures and hazard recognition but also how the robot works. This enables workers to understand behavior and respond to unexpected situations.
By 2028, 330-pound walking machines will move through factory floors in Savannah. The question is whether we’ll have established adequate protections before accidents happen, or whether we’ll learn from preventable injuries after the fact.
Our articles make government information more accessible. Please consult a qualified professional for financial, legal, or health advice specific to your circumstances.