Autonomous ships are designed to navigate, adjust speed, avoid collisions, and perform other functions with less direct human involvement. Some are controlled by workers in remote operations centers, while others can perform certain tasks using software, sensors, cameras, radar, and artificial intelligence.
These technologies could reduce accidents caused by fatigue, distraction, and other forms of human error. They can also introduce new risks involving defective software, failed communications, inaccurate sensor data, cyberattacks, and confusion over who is responsible when something goes wrong.
The regulatory landscape reached a major turning point in May 2026, when the International Maritime Organization adopted the first global safety code for autonomous ships. The non-mandatory code took effect on July 1, 2026, and places human oversight, risk assessment, cybersecurity, remote operations, and accountability at the center of autonomous vessel operations.
The arrival of autonomous technology does not eliminate legal responsibility. Instead, it may distribute responsibility among the vessel owner, operator, master, employer, remote operator, technology manufacturer, software developer, and other companies involved in the vessel’s operation.
What Is an Autonomous Ship?
The International Maritime Organization uses the term Maritime Autonomous Surface Ship, or MASS, for a ship that can operate independently of human interaction to some degree.
Autonomy does not always mean a vessel is completely uncrewed. An autonomous ship may have:
- A traditional crew assisted by automated navigation systems
- A reduced crew supported by shoreside operators
- A remote master or operator controlling the vessel from land
- Autonomous systems that make certain decisions without immediate human input
- The ability to operate without crew during approved portions of a voyage
A ship does not become a MASS simply because it has advanced automation. Under the new framework, its autonomous or remote functions must replace or support tasks normally performed by onboard crew, and the vessel must complete an approval process and hold a valid MASS Safety Certificate.
The IMO also emphasizes that autonomous ships receive no special privilege over conventional vessels. They must operate alongside crewed vessels and comply with existing international safety and navigation requirements.
Maritime companies are already using artificial intelligence and other technologies to improve safety. Autonomous navigation represents a more advanced stage of that trend because software may take over decisions that were traditionally made by a captain, pilot, or bridge crew.
What Does the New MASS Code Require?
The International Code of Safety for Maritime Autonomous Surface Ships establishes a goal-based framework for the design, approval, certification, and operation of autonomous and remotely controlled commercial ships.
The code currently applies to cargo ships covered by Chapter I of the International Convention for the Safety of Life at Sea. These are generally cargo ships of 500 gross tons or more operating on international voyages. Governments are encouraged to apply the principles to smaller autonomous ships when practical.
Important areas addressed by the code include:
- Risk assessment
- Navigation
- Software design
- Connectivity
- Remote operations centers
- Alert management
- Fire detection and protection
- Machinery and electrical systems
- Cybersecurity
- Search and rescue
- Training and watchkeeping
- Safe operating limits
- Emergency and fallback procedures
The code calls for operators to define the vessel’s operational context, including the weather, visibility, sea state, water depth, traffic, and other conditions under which its autonomous systems can function safely. There must also be a plan for responding when conditions move outside those limits.
Because the code is currently non-mandatory, it does not create one universal civil-liability rule for every autonomous ship accident. It is intended to help governments and the maritime industry gain experience before a mandatory code is developed. The IMO currently expects to adopt a mandatory MASS Code by July 2030 for entry into force in 2032.
Is There Still a Master on an Autonomous Ship?
Yes. The MASS Code maintains the concept of a human master, even when that person is not physically aboard the vessel.
The master retains overall responsibility for the ship at all times. The master must also have a way to intervene when necessary, whether control is exercised from the bridge, a shoreside remote operations center, or another approved location.
A remote operations center may contain masters, navigators, engineers, system technicians, and other personnel who monitor or control one or more vessel functions. The code calls for these centers to be assessed and certified and to operate under a safety management system.
Keeping a human master in the chain of command helps preserve accountability, but it does not answer every liability question. An investigation may still need to determine:
- Whether the master received accurate information
- Whether the master had enough time to intervene
- Whether communications with the ship were available
- Whether the interface clearly communicated the danger
- Whether the master was responsible for too many functions or vessels
- Whether the autonomous system followed or overrode a human command
- Whether the company provided adequate staffing and training
Are Autonomous Ships Legal in the United States?
Autonomous and remotely operated vessels are not categorically prohibited in the United States, but they remain subject to existing vessel safety, inspection, navigation, and manning laws.
The U.S. Government Accountability Office has reported that the Coast Guard regulates autonomous vessels through existing laws and regulations. The agency also noted that many federal statutes were written with the assumption that crew members would be aboard and in control of a ship. Existing minimum-crewing laws may therefore prevent some U.S. vessels from operating without crew, even when the technology is capable of doing so.
In June 2026, the Coast Guard published new guidance for unmanned, autonomous, and remote-control operations. It established review procedures for autonomous systems used on uninspected vessels, passenger vessels, and vessels subject to Coast Guard inspection.
Approval to test or operate one autonomous system does not necessarily authorize every route, operating condition, or use. Coast Guard officials may consider the vessel’s design, intended operation, inspection status, navigation area, passenger capacity, level of human control, and proposed safety measures.
Who May Be Liable for an Autonomous Ship Accident?
Liability depends on why the accident happened, who controlled the relevant function, where it occurred, who was injured, and which companies were responsible for the vessel and its technology.
More than one party may share responsibility.
The Vessel Owner or Operator
The vessel owner or operating company may be liable when an accident results from unsafe deployment, operation, maintenance, or supervision of an autonomous system.
Examples could include:
- Using an autonomous system outside its approved operating limits
- Failing to maintain cameras, radar, sensors, or communications equipment
- Ignoring repeated software warnings or system faults
- Operating without an adequate fallback system
- Failing to provide sufficient remote operators
- Assigning one operator more vessels than could be monitored safely
- Failing to perform required risk assessments
- Continuing a voyage after losing essential connectivity
- Failing to update known cybersecurity vulnerabilities
- Using technology that had not been properly tested or approved
The owner cannot necessarily avoid responsibility by arguing that the computer made the decision. A company that selects, configures, maintains, and deploys the system may still be responsible for foreseeable failures and unsafe operating practices.
The Vessel Master
The master’s conduct may be examined when the master had authority to intervene but failed to respond appropriately.
An investigation may consider whether the master properly monitored the vessel, understood its automated systems, maintained situational awareness, and acted when the vessel approached an unsafe condition.
The MASS Code’s requirement that a human master retain overall responsibility means the master’s location alone does not end the inquiry. A master operating from shore may still be responsible for the vessel, although the precise legal consequences will depend on the applicable laws and facts.
The owner or operator may also be responsible for the master’s actions when the master was acting within the scope of employment.
A Remote Operator or Remote Operations Center
A remote operator may control navigation, propulsion, machinery, cargo systems, or emergency functions from a shoreside facility.
Potential failures may include:
- Misinterpreting sensor data
- Failing to maintain a proper watch
- Delaying intervention
- Selecting the wrong vessel from a control interface
- Transmitting an unsafe command
- Failing to notify the master of a developing problem
- Overlooking an alarm
- Losing awareness during a handoff between operators
- Operating while fatigued or distracted
Liability may rest primarily with the company that employed, trained, scheduled, and supervised the operator. In some circumstances, an operator’s individual conduct may also be examined.
The Employer of an Injured Seaman
A qualifying seaman injured during an autonomous vessel accident may have rights under the Jones Act.
Under 46 U.S.C. § 30104, a seaman injured in the course of employment may bring a civil action against the employer. Employer negligence involving autonomous technology could include inadequate training, insufficient staffing, unsafe operating procedures, poor maintenance, or the decision to use a system despite known defects.
The employer’s negligence does not have to involve a traditional navigation mistake. It could involve decisions made months before the accident, such as selecting an unsafe system, failing to test it, ignoring warnings from crew members, or failing to prepare workers for an emergency transition from autonomous to manual control.
An injured seaman may also be entitled to maintenance and cure, a separate maritime benefit that generally does not require proof that the employer caused the injury.
The Vessel Owner Under the Doctrine of Unseaworthiness
A seaman may also have a claim under the doctrine of unseaworthiness.
A vessel owner has a duty to provide a vessel, equipment, and crew reasonably fit for their intended use. The duty does not require a perfect or accident-free ship, but the vessel must be reasonably suitable for its intended service.
A malfunctioning autonomous navigation system, unreliable sensor array, inadequate backup control, or incompetent remote operating team could potentially support an unseaworthiness claim. Whether software or a remotely operated system legally qualifies as an unseaworthy vessel component would depend on the system’s function, integration with the vessel, and role in causing the injury.
Because autonomous-vessel case law is still developing, courts will likely apply established maritime principles to unfamiliar combinations of software, shoreside personnel, and onboard machinery.
The Technology Manufacturer or Software Developer
A company that designs or manufactures an autonomous navigation system may be responsible if a defect causes physical injury or damages property other than the defective product itself.
Potential defects may involve:
- Unsafe system design
- Manufacturing flaws
- Inadequate testing
- Defective sensors or control units
- Incorrect navigation data
- Failure to warn operators about known limitations
- Software that issues unsafe commands
- Inadequate cybersecurity protections
- Faulty integration between hardware and software
The Supreme Court has recognized product-liability principles as part of general maritime law. However, it has also held that a commercial party generally cannot use a maritime product-liability claim to recover purely economic losses when a defective product damages only itself. Claims involving bodily injury or damage to other property are treated differently.
Learn more about claims involving defective maritime equipment and products.
Software can make these cases more complicated. The shipbuilder may blame the software company, the software company may blame a sensor manufacturer, and all parties may argue that the operator failed to use the system correctly.
Maintenance, Integration, and Communications Contractors
Autonomous vessels may rely on several outside contractors to install, update, monitor, and maintain their systems.
A contractor could share responsibility if it:
- Installed sensors incorrectly
- Failed to test the control system
- Introduced an unsafe software update
- Misconfigured an alarm or operating limit
- Failed to repair a known defect
- Provided unreliable connectivity
- Certified a system without adequate testing
- Failed to communicate a known risk to the vessel operator
Contracts between the parties may determine who was responsible for particular systems, but contractual language does not always prevent an injured person from pursuing an otherwise valid maritime claim.
What Types of Failures Can Cause Autonomous Ship Accidents?
Autonomous ships depend on several systems working together. A failure in one component can affect the entire vessel.
Sensor Failure
Autonomous navigation systems use radar, cameras, lidar, sonar, the Automatic Identification System, satellite navigation, and environmental sensors to understand their surroundings.
A dirty camera lens, obstructed sensor, radar fault, incorrect GPS location, or failure to recognize a small vessel could prevent the system from detecting a collision risk.
Connectivity Loss
A remotely operated vessel may depend on satellite, radio, cellular, or other connections to communicate with its operations center.
Connectivity can be disrupted by weather, equipment failure, interference, geographic limitations, or deliberate attack. A safe system should have defined procedures for slowing, stopping, changing modes, or otherwise reaching a safe condition when communication is lost.
Software and Decision-Making Errors
Software may correctly identify an object but choose an unsafe response. It may also make decisions based on inaccurate assumptions, outdated charts, incomplete data, or a scenario that was not adequately represented during testing.
Unlike an ordinary equipment failure, a software error may occur across an entire fleet if several ships use the same system or update.
Human-Machine Handoff Failures
A critical risk can arise when control passes between an autonomous system and a human operator.
An operator may not immediately understand:
- Why control was transferred
- Which hazards the system detected
- What commands the system already issued
- How much time remains before a collision
- Whether the vessel will continue following its previous course
- Which systems are still operating autonomously
An emergency handoff is not safe merely because a human is technically able to take control. The person must receive understandable information and enough time to respond.
Cyberattacks
Autonomous and remotely operated vessels may be exposed to attacks targeting navigation, propulsion, communications, cargo systems, or operational data.
The GAO has identified growing cybersecurity risks throughout the U.S. maritime transportation system, including threats from foreign governments and transnational criminal organizations. The Coast Guard’s current cybersecurity regulations establish minimum requirements for certain U.S.-flagged vessels and regulated maritime facilities.
A cyberattack does not automatically make the vessel owner negligent. However, liability questions may arise if a company failed to install available security updates, ignored warnings, used weak access controls, lacked an incident-response plan, or continued operating after learning that a critical system had been compromised.
Read more about cybersecurity risks in the maritime industry.
Do Autonomous Ships Still Have to Follow Navigation Rules?
Yes. Autonomous and uncrewed vessels remain subject to applicable collision-avoidance and navigation rules.
The Coast Guard explains that the Navigation Rules apply broadly to watercraft used or capable of being used for transportation. The rules do not give an autonomous vessel priority over a crewed vessel.
Rule 5 requires every vessel to maintain a proper lookout by sight, hearing, and all available means appropriate to the circumstances. The Coast Guard specifically states that this responsibility applies to unmanned craft as well as crewed vessels.
For an autonomous ship, maintaining a lookout may involve cameras, microphones, radar, software, and human monitoring from a remote center. After a vessel collision, investigators may examine whether those systems collectively provided an adequate lookout.
Other relevant duties may include:
- Traveling at a safe speed
- Assessing collision risk
- Taking timely action to avoid a collision
- Accounting for restricted visibility
- Using proper lights and sound signals
- Following narrow-channel and traffic-separation rules
- Exercising ordinary seamanship and reasonable precautions
A vessel cannot excuse an unsafe maneuver simply by showing that its software behaved as programmed.
Are Remote Operators Protected by the Jones Act?
A shoreside remote operator does not automatically qualify as a Jones Act seaman merely because their work contributes to a vessel’s operation.
The Supreme Court’s seaman-status test requires a worker to contribute to a vessel’s function or mission and have a connection to a vessel or identifiable fleet that is substantial in both duration and nature. The test is intended to distinguish sea-based crew members from land-based workers and focuses partly on regular exposure to the risks of maritime service.
A remote operator who works entirely from a land-based control center may have difficulty meeting that test. A worker who also spends substantial time aboard the vessel, participates in voyages, or regularly transitions between onboard and remote duties may present a different case.
Workers who do not qualify as seamen may have rights under state workers’ compensation laws, the Longshore and Harbor Workers’ Compensation Act, or another employment or maritime law. Coverage depends on the worker’s actual duties and location, not simply the job title “remote operator.”
What Evidence Is Important After an Autonomous Ship Accident?
Traditional maritime investigations examine witness accounts, weather, navigation decisions, maintenance records, crew training, and the vessel’s physical condition. Autonomous ship investigations may require much more electronic evidence.
Important records may include:
- Voyage data recorder information
- Radar and Automatic Identification System data
- Camera and sensor recordings
- Remote operations center communications
- Commands sent to and from the vessel
- Software versions and update histories
- System alerts and error logs
- Cybersecurity and access records
- Human override attempts
- Maintenance and inspection records
- Risk assessments
- Approved operating limits
- Safety-management procedures
- Remote operator schedules and fatigue records
- Training and certification records
- Contracts identifying responsibility for each system
This information can help determine whether the accident resulted from defective technology, poor maintenance, operator error, inadequate training, an unsafe company policy, or several contributing failures.
Electronic records can be overwritten, updated, or distributed among multiple companies. Preserving evidence quickly may therefore be especially important after an autonomous vessel accident.
Frequently Asked Questions About Autonomous Ship Liability
Does the MASS Code decide who must pay after an accident?
No. The MASS Code establishes safety, approval, and operational expectations, but it does not provide one civil-liability formula for every accident. Liability will depend on applicable national and international law and the specific cause of the incident.
Is the vessel owner liable whenever an autonomous system fails?
Not necessarily. An investigation must determine why the system failed, whether the failure was foreseeable, who was responsible for the system, and whether reasonable safety measures were taken.
The owner may share liability with a manufacturer, operator, contractor, employer, or another vessel.
Can an artificial intelligence system be sued?
Legal claims are generally brought against the people and organizations that design, own, operate, maintain, or deploy the system rather than against the software itself.
The difficult question is determining which person or company controlled the dangerous decision and had the ability to prevent it.
Can an autonomous vessel be unseaworthy?
Potentially. A vessel may be unseaworthy when it, its equipment, or its crew is not reasonably fit for the vessel’s intended purpose.
An unreliable autonomous navigation system, inadequate backup control, or insufficiently trained operating team could support an unseaworthiness argument when the condition contributes to a seaman’s injury.
What happens if an autonomous ship collides with a crewed vessel?
Investigators may examine the conduct of both vessels. Responsibility could be divided if both failed to maintain a proper lookout, travel at a safe speed, communicate, or take appropriate collision-avoidance action.
The autonomous vessel’s owner cannot necessarily avoid liability by blaming its software, and the crewed vessel is not automatically free from fault because the other ship was autonomous.
Who is responsible when a cyberattack causes the accident?
The attacker may face criminal and civil consequences if identified. Other parties may also be examined.
A vessel owner, operator, software provider, or contractor could potentially share responsibility if inadequate security practices allowed a foreseeable attack to cause harm. Each case requires an analysis of the attack, the available protections, and the parties’ conduct.
Getting Help After an Autonomous Ship Accident
Autonomous vessel accidents may involve familiar maritime laws but unusually complex evidence. Several companies may control different parts of the vessel, and important information may be stored aboard the ship, at a remote operations center, and on systems maintained by third-party technology providers.
Injured workers and passengers should seek medical care and report the accident promptly. When possible, they should also preserve photographs, communications, witness information, medical records, and details about the vessel and technology involved.
A maritime lawyer can investigate whether responsibility rests with the employer, vessel owner, operator, remote operations center, technology manufacturer, contractor, or multiple parties. An attorney can also determine whether the Jones Act, unseaworthiness, maintenance and cure, product-liability law, or another maritime remedy applies.
As autonomous shipping expands, the technology used to operate vessels will change. The basic expectation that companies protect workers, passengers, and other people on the water will remain.