AI companies may be coming, but AI risks are already here

Key takeaways

  • As Delaware begins experimenting with a new business concept — artificial intelligence companies — insurance buyers should already be addressing their present-day AI exposures.

  • AI-related securities claims are accelerating, necessitating clear communication and defensible disclosure policies about a company’s use of AI.

  • Executives should strengthen AI governance, vendor controls, and contractual provisions, with an understanding of ownership, accountability, and responsibility for AI risks.

It almost feels like a scene from a science fiction movie: an entire company that operates and governs itself solely with autonomous artificial intelligence (AI).

No humans at the helm. Just AI agents owning assets, running the till, and executing corporate strategies.

The concept is approaching reality in Delaware.

Delaware is testing a novel corporate structure to create a legal framework for autonomous businesses: the artificial intelligence company (AIC). Initial activity includes a regulatory sandbox, in which select AICs can operate during a 30-month pilot program, after which state lawmakers may consider establishing a permanent legal entity for AICs.

The idea has garnered broad discussion, including from members of Congress. In October, James Comer, a Kentucky Republican who chairs the House Committee on Oversight and Government Reform, criticized the AIC proposal’s liability shield (opens a new window) as a means to protect AI companies and their investors from responsibility for harmful outcomes.

There are already examples of companies where AI has operational autonomy, although the concept of AI as the ultimate corporate decision-maker remains experimental. Regardless of whether AICs become viable business models, they raise unresolved questions about accountability, liability, and regulatory control. The incredible rate of AI adoption in traditional human-run corporations and the scale of its deployment introduce novel legal risks and insurance considerations.

As more companies embed AI in their businesses, discuss it publicly, and automate certain functions — all arguably incremental steps toward a full-blown AIC — corporate leaders should be aware of the growing AI-related risks to their businesses.

How AI poses risks to top executives

For publicly traded companies, securities class-action lawsuits are one of the most prevalent AI-related risks today.

Cornerstone Research reports (opens a new window) that nearly as many AI-related securities lawsuits were filed in the first half of 2026 as all of 2025.

Simply discussing AI poses a liability risk for corporate boards to consider. Shareholders can be quick to bring claims if directors and officers overstate AI’s benefits to the business or the role of AI in a product or service, a concept known as “AI-washing”, or if they understate or conceal its risks.

Similarly, shareholders can allege that directors and officers concealed the company’s competitive disadvantage resulting from a third party’s use of AI that diverted resources or customers. A related exposure can occur if a company’s AI adoption disrupts its business model more than it helps, or results in increased competitive pressure.

Other claims can include allegations that the company made inaccurate or misleading statements about an outside party, such as an independent researcher, validating the company’s use of an AI tool.

Regulatory agencies have similarly initiated enforcement actions against companies alleged to have engaged in deceptive puffery about the impact of AI on their businesses. Specifically, the SEC has focused on various scenarios, such as companies making misleading statements regarding AI use, the role it plays, its performance, or its benefits. And the Federal Trade Commission (FTC), guided by its goal of protecting consumers, has taken action in instances where companies have overstated AI capabilities or performance without adequate substantiation.

Such claims are likely to trigger directors and officers liability (D&O) insurance, which can cover defense costs and indemnify executives and the companies they serve, for allegations of misstatements, mismanagement, or professional malfeasance.

Implications for vendor and supply chain management

Companies that rely on outside vendors to supply AI-related tools or services need to give deeper consideration to the AI supply chain and assignment of responsibility when the technology fails or underdelivers.

The failure of an AI vendor’s product does not absolve its users of liability.

For example, if a company uses a vendor’s AI application to screen candidates, and that AI model has defects or biases that result in discriminatory candidate selections, both companies could face liability. Depending on the circumstances, an employment practices liability (EPL) policy, D&O, technology errors and omissions liability (Tech E&O) policy, or professional liability policy may respond.

Where do companies decide to incorporate?

For decades, Delaware has reigned as the corporate home of choice for publicly traded companies in the U.S., due to a well-established body of law designed to protect businesses and shareholders, paired with a sophisticated judiciary. In recent years, however, Delaware has come under pressure from other states like Texas, Nevada, and Wyoming that would like to usurp its status as the preferred locale for corporate formation and jurisprudence.

The most visible example occurred when Elon Musk incorporated SpaceX in Texas and encouraged others to do the same. (opens a new window)

Incorporating elsewhere carries risks, though, particularly because companies lose a measure of predictability from Delaware that other states lack. Delaware’s first move on AICs may be a calculated approach to retain its status given the technology to establish autonomous companies is already emerging.

Advancement of the AIC proposal could widen that gap.

Practical considerations

Executives should work with legal counsel and risk managers to develop policies and guidelines for how companies discuss and disclose their use of AI and the associated risks. But even well-designed policies and effective training programs may still not prevent a statement or disclosure from triggering a securities claim. That is why it’s vital that companies coordinate with experienced insurance brokers to design D&O programs that can respond if claims arise.

Companies should also develop a robust governance framework and establish processes to map and understand their AI supply chain and vendor risks. Alignment with a framework, such as the National Institute of Standards and Technology’s AI Risk Management Framework, can assist in these efforts.

Contracts and agreements with AI-related vendors and customers affected by the technology should clearly define ownership, responsibility, and indemnification rights and obligations for errors, hallucinations, outages, and other AI-driven risks.

E&O, tech E&O, cyber, and EPL policies can be put in place and crafted to respond to the residual risk that remains. AICs may come to fruition in the future, and when they do, they will present new and complex risk issues, including the possible need for dedicated AI-specific insurance products. But companies can address their real and present-day AI exposures now.