U.S. Navy Unmanned Vessel Reportedly Capsized Support Boat After Receiving Inadvertent Command

During U.S. Navy testing at Channel Islands Harbor Marina on June 23, 2025, an unmanned vessel under tow reportedly received an inadvertent command and accelerated, capsizing the support boat and throwing its captain into the water. Emergency responders reportedly recovered the captain, who declined medical attention. The incident was reportedly attributed to software failures and human error, including communication breakdowns between onboard systems and external autonomy software.

Risk classification

• Primary risk domain: 7 AI system safety, failures, & limitations
• Primary risk subdomain: 7.3 Lack of capability or robustness

The unmanned vessel failed to operate safely and robustly, executing an inadvertent command that turned its engine on while being towed, leading to a physical accident.

Causal factors

• Entity: Other
• Intent: Unintentional
• Timing: Post-deployment

The incident was caused by an inadvertent command to an unmanned vessel during an operational test, though it remains ambiguous whether this was due to human error or a technical malfunction.

EU AI Act risk tier

• Risk tier: 2 High Risk

Risk Level 2: High Risk. The system is an unmanned maritime vessel operating in a harbor, which represents an AI system with significant implications for physical safety and critical transport infrastructure.

AI system and alleged parties

• AI system: unmanned surface vessel
• AI purpose: Autonomous Drones; Navigation Assistant
• Behaviour type: Autonomous
• Alleged developer: Unmanned surface vessel manufacturers, Defense technology companies, BlackSea Technologies, Autonomous maritime systems developers
• Alleged deployer: United States Navy, United States Armed Forces, Naval forces, Military organizations, Military autonomous systems operators
• Alleged harmed parties: Watercraft, Vessel operators, Support boats, Maritime workers

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Minor. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Negligible
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Minor, indirect Negligible
• Financial: direct Negligible, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Negligible, indirect Negligible
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Negligible, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Property

Reported: The report explicitly describes property damage in the form of a capsized support boat.

Directly caused: The support tugboat was overturned and capsized by the accelerating unmanned vehicle.

Inferred additional harm: The capsized tugboat likely suffered water damage to its engine, electronics, and structure, though the exact financial value of this damage is not specified.

People affected

• Occurrences reported: 1
• People reportedly harmed: 1
• People reportedly exposed: 1

Potential causes

Management

• Inadequate Pre-Test Risk Mitigation: Safety procedures were only updated via a stand-down after the incident.

Technology

• Inadvertent Command Reception: The unmanned vehicle received an unintended command that turned the engine on.
• Sudden Vehicle Acceleration: The vehicle accelerated past and capsized the support boat after engine start.

Data Inputs

• Inadvertent Control Signal: An unintended input command was transmitted to or processed by the vehicle.

Human Factors

• Unclear Command Source: It is undetermined if a human operator inadvertently triggered the command.

Process and Methods

• Deficient Towing Protocols: Safety procedures during towing did not prevent accidental engine activation.

Information quality

• Classification confidence: Medium
• Reason for confidence: The incident details regarding the capsizing and rescue are clearly reported by official spokespersons. However, the exact cause of the 'inadvertent command' (whether human error or technical malfunction) remains unconfirmed, which limits our ability to fully classify the causal entity.
• Ambiguities identified: Whether the 'inadvertent command' originated from a human operator error or a software/technical malfunction.
• Alternative interpretations: The incident could be interpreted purely as a human operational error during towing rather than an AI system failure, depending on the origin of the inadvertent command.
• Missing information: The source and nature of the 'inadvertent command' that activated the engine, and the specific model of the unmanned vessel involved.

A U.S. Navy unmanned surface vessel test resulted in a localized physical accident when an inadvertent command caused the drone to accelerate and capsize its support tugboat. The incident, which led to a temporary safety stand-down, represents a minor physical safety failure in military R&D rather than an adversarial national security threat.

• Overall national security impact: Minor
• Scope: Single nation
• Primary target: United States
• Alleged perpetrator: Unknown

Threat characteristics

• Imminence: Long-term. The immediate incident is resolved and a safety stand-down was initiated, representing an ongoing system safety and development concern rather than an active crisis.
• Autonomy: Human-supervised. The vessel is an unmanned surface vehicle designed for autonomous operations but was being towed and tested under human supervision.
• Novelty: Established threat. Accidents and software glitches during experimental military drone testing are established risks with existing precedents.

Impact by dimension

• Physical security: Minor. The unmanned vessel received an inadvertent command, causing it to accelerate and capsize a support tugboat, throwing the captain into the water. This represents a minor physical safety incident during a controlled military test.
• Information security: Negligible. No evidence of intelligence compromise, classified data theft, or information warfare activities in the reported incident.
• Sovereignty: Negligible. The incident occurred during a domestic military training exercise in U.S. waters with no foreign involvement or threat to government operations.
• Economic security: Minor. Minor property damage to the capsized support tugboat and a temporary safety pause in Navy testing operations, but no strategic technology theft or systemic economic threat.
• Societal stability: Negligible. No societal stability, mass surveillance, or human rights implications are associated with this localized military testing accident.