Physical artificial intelligence, which includes technologies like self-driving cars and humanoid robots, is set to change the insurance world in big ways. Chris Raimondo, the insurance consulting leader for the Americas at EY, says this shift might have a bigger impact on insurance than the digital AI tools many companies use today.
Spending on AI is growing quickly. The International Data Corporation expects global investment in AI to top $500 billion by 2027, with more of that money going to physical AI applications such as robotics and autonomous vehicles. Goldman Sachs also predicts the humanoid robotics market alone could reach $150 billion by 2035, fueled by labor shortages and advances in machine learning.
Raimondo points out that these physical AI systems will change how liability is handled in insurance. Traditionally, liability in areas like car insurance has focused on the human driver. But with autonomous systems, responsibility may spread across different parties — from the vehicle maker to the software provider and the owner of the vehicle, whether they are individuals or companies.
This change is already underway. The US National Highway Traffic Safety Administration reports that advanced driver-assistance and self-driving features are being used in millions of vehicles. Some manufacturers have logged billions of autonomous miles during testing and limited commercial use. Insurers will need to consider risks connected not just to drivers but also to the vehicle’s software, sensors, and the data it produces.
Insurance companies are beginning to adapt. For example, insurtech firm Lemonade has launched a usage-based policy that changes the price depending on whether a car is being driven by a human or in autonomous mode. This points toward a new kind of insurance where premiums are adjusted in real time based on data from AI-enabled systems.
It’s not just the vehicles that need coverage. Raimondo says insurance for the supporting infrastructure like charging stations, data centers, and smart roads is gaining attention too. Claims processing is also expected to evolve. Instead of relying mainly on human investigation, insurers will use sensor data, logs, and software diagnostics to understand what went wrong.
Although generative and agentic AI have grabbed headlines recently, their effects on insurance have mostly improved internal processes like underwriting and fraud detection. Physical AI goes further, affecting what insurance actually covers.
Raimondo emphasizes that the whole insurance industry will have to change—not just products but also the entire business model. While this shift is in the early stages, growing investment and adoption are pushing insurers to rethink how they work. The speed of this change will vary by region, but with more technology on the roads and in robots, the pressure to adapt will only increase.