From Generation to Action

The first wave of generative AI helped people create content. The next wave is extending AI beyond content generation: autonomous AI systems can perceive their environments, make decisions and execute actions in both digital and physical environments.

These systems range from software that can issue refunds, place orders or coordinate workflows to autonomous physical systems that transport goods, operate machinery or interact with people.

This transition is already accelerating. The share of organisations reporting the use of AI increased from 55% in 2023 to 78% in 2024[1]. Looking ahead, 86% of employers expect AI and information-processing technologies to transform their business by 2030, while 58% expect robots and autonomous systems to do the same[2].

These risks are not merely theoretical. Several organisations have already experienced significant financial and legal consequences from decisions, outputs and actions involving AI systems.

When AI Creates Real-World Loss

In 2025, a federal jury found Tesla partly responsible for a 2019 crash involving its Autopilot driver-assistance system. The crash killed one person and seriously injured another. The jury awarded approximately $243 million in compensatory and punitive damages, and the verdict was upheld by a federal judge in February 2026[3].

The driver was also found responsible for the crash. The case therefore illustrates a central problem created by autonomous AI systems: liability may be divided between the system's developer, its operator and the company that placed it into use.

In 2024, fraudsters used AI-generated video and audio to impersonate senior employees of engineering firm Arup. After joining a video call with what appeared to be several colleagues, an employee authorised transfers totalling approximately $25 million[4].

The Coverage Gap

AI-related losses do not fit neatly within a single traditional insurance category. A data breach may fall under cyber insurance, an incorrect professional decision under errors and omissions coverage, fraud under a crime policy, and physical injury under general or product liability. This patchwork creates uncertainty when an autonomous system contributes to several types of loss at once. Some policies remain silent on AI, while insurers are also introducing targeted exclusions and endorsements. Businesses therefore cannot assume that an existing policy will respond unless AI-related actions and losses are addressed explicitly[5][6].

As Authority Grows, Protection Must Evolve

Autonomous AI systems will increasingly influence transactions, business operations and physical environments. Their benefits may be substantial, but greater autonomy also increases the speed and scale at which mistakes can create damage. Financial protection must therefore evolve alongside the technology. Companies need coverage that clearly defines which AI-related losses are insured, while insurers need access to the system data and controls required to understand the risk. The central question is no longer only what an AI system can produce, but what it is authorised to do, how much damage it can cause and who pays when it fails.