Physical AI: Implications for risk managers, insurers, and adjusters

How physical AI is transforming operations across industries

Physical artificial intelligence (AI) is rapidly transforming operational environments across logistics, warehousing, and transportation. Autonomous warehouse systems and driverless trucking fleets are now being deployed at scale, fundamentally shifting how risk is created, concentrated, and transferred. This evolution introduces a blended risk profile combining physical assets, software-driven decision-making, and system-wide interconnectivity.

The transition to physical AI is not speculative; it is already being implemented across sectors, often through multi-year transformation programs.

Major retailers and logistics companies are converting distribution centers into autonomous operations. Autonomous trucking is already being tested and deployed in real-world environments. Robotics are being introduced across nearly every sector of the economy, particularly in industries such as mining, manufacturing, distribution and healthcare. In many cases, organizations are upgrading one facility or one location or system at a time, gradually shifting their risk exposure as each transition is completed.

This incremental adoption creates a unique challenge: a single organization may simultaneously operate traditional and autonomous environments, each with very different risk characteristics. As more facilities convert, more exposure shifts into unfamiliar territory.

Autonomous warehouse systems and factory robotics are evolving beyond traditional programmable machines with defined boundaries.

Physical AI increasingly includes agentic AI and autonomous, large language model-driven robots. These systems are designed to interpret their surroundings, make decisions, and adapt their behavior based on changing conditions. They are not simply executing instructions; they are, in effect, making complex decisions in dynamic, evolving environments.

Physical AI is already being deployed at scale across manufacturing, warehousing, and transportation. Centralized AI platforms can coordinate hundreds of autonomous robots within a single operation, enabling higher-density facilities, continuous production, and ongoing optimization. At the same time, autonomous trucking technologies are moving from pilot programs to commercial deployment, with driverless vehicles operating in designated freight corridors and accumulating millions of miles of operational experience. Industry forecasts indicate rapid fleet expansion in the coming years.

How physical AI is changing risk exposure and loss profiles

The adoption of physical AI is shifting risk in several critical ways:

• Autonomous environments often condense more operations into smaller spaces, increasing exposure density.
• From isolated equipment failures to system-wide events impacting entire facilities or fleets
• From human error to algorithmic and autonomous decision making
• From traditional warehouse exposures to concentrated, high-value asset environments
• From primarily human injury exposure to specialized equipment and asset damage

As these assets become more expensive and safer, the risk profile is shifting. In some areas, loss frequency is decreasing, but when incidents do occur, they are more severe, take longer to restore, involve more complex subrogation investigations, and lead to significantly higher business interruption and extra expense exposure.

How claims and liability are evolving in autonomous environments

As physical AI blurs the line between physical and digital systems, claims investigation must evolve accordingly. Determining cause of loss now requires more than a physical inspection. 

Autonomous systems introduce new claims challenges, including:

• Failure, origin and cause investigations across interconnected systems
• Increased asset values and specialized (skilled labor) and equipment restoration requirements
• Data-driven incident analysis requiring system logs and telemetry review
• Increased business interruption timelines due to product lead time replacement and system validation and recalibration or resequencing.
• Increased business interruption exposure as automation reduces operating costs and drives higher profit margins.
• The loss or damage of entire fleets of automated equipment and operational systems.

Claims investigations increasingly require multidisciplinary expertise, including forensic robotics engineering, advanced data analysis, and agentic software evaluation.

Autonomous trucking liability is evolving rapidly from driver-centric to shared responsibility across manufacturers, AI developers, and fleet operators. Real-time data capture enhances incident reconstruction but introduces additional complexity in determining causation and responsibility if a system failure occurs.

Insurance models are adapting to include blended coverage across auto liability, product liability, and cyber risk (vehicle hacking).

Regulatory momentum and the future of autonomous systems

Since 2017, the U.S. Department of Transportation and leading autonomous trucking companies have focused on a shared priority: safety. This has been the number one priority for the government, owners, operators, consumers and all the other drivers on the road. This is understandable given that there are approximately 500,000 vehicle accidents a year involving trucks. From 2021 – 2023 large truck crashes led to more than 15,000 deaths. However, according to the DOT, over 90% of the crashes are caused by human error. To date, early operational data suggests autonomous driving technologies have the potential to reduce crashes, injuries and fatalities.

The H.R. 4661 – AMERICA DRIVES Act proposes a federal framework for autonomous commercial trucking, including the operation of Level 4 and Level 5 vehicles without human drivers across state lines. Proposed federal frameworks and ongoing regulatory discussions may enable broader interstate deployment of autonomous trucking, though state-level adoption remains uneven and continues to evolve. Several states (including Arizona, Texas, Nevada, Florida, Oklahoma, and Ohio) are already supporting testing and limited pilot programs, providing valuable operational experience as the technology matures.

What insurers and adjusters should do now

The rapid adoption of physical AI requires insurers and adjusters to begin adapting today.

As physical AI continues to evolve, organizations cannot wait for full adoption to act. Several steps can help position them more effectively for this transition:

• Updating underwriting models and analyzing each risk at renewal
• Updating the insured’s current implementation of Physical AI to reflect system-level risks and Physical AI interconnectivity
• Reevaluating coverage structures for business interruption and high-value assets
• Developing or updating multidisciplinary claims capabilities
• Monitoring regulatory developments impacting autonomous operations

Organizations that proactively adapt will be better positioned to manage emerging exposures and maintain claims defensibility.

Preparing for a more complex, system-driven risk landscape

Physical AI represents a fundamental shift from traditional, linear risk environments to complex, adaptive systems. Loss events are increasingly driven by how systems interpret, decide, and interact at scale. For insurers and adjusters, success will depend on the ability to evaluate both the physical and digital dimensions of loss, develop multidisciplinary expertise, and respond effectively to a risk landscape that is increasingly interconnected, automated and data-driven.

Sources cited:

https://www.fmcsa.dot.gov/safety/data-and-statistics/large-truck-and-bus-crash-facts
https://crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/813717.pdf