The rise of AI: implications for D&O insurance wordings
10/21/2025
Jonathan Hopkins of DAC Beachcroft and Zurich’s Jack O’Neill tackle the opportunities and challenges for brokers. Clients are asking about the impact of artificial intelligence on directors’ and officers’ insurance policies in a fast-changing landscape. Are existing policies, broadly, fit for purpose?
This article was originally published on Insurance Age. You can read the article here.
According to Amazon Web Services, at least one business adopts AI every 60 seconds in the UK; with more than 52% of businesses now using AI. It's an increasingly hot topic for management teams who want to maximise operational efficiencies by incorporating AI, whilst at the same time managing associated risks.
This includes risks to the directors themselves, who may be held accountable for management failings directly linked to how AI is implemented and overseen. For brokers, this presents both opportunities and new challenges: clients may be asking how their Directors & Officers (D&O) insurance responds to a claim or loss relating to AI, and what protections are in place for senior leaders navigating this fast-changing landscape.
Here we look at whether existing D&O insurance policies are, broadly, fit for purpose.
What is AI?
To manage AI risks, it's important for companies to understand just what “AI” actually is. “AI” is an umbrella term that covers a wide range of systems, including:
- Basic machine learning uses repetitive tasks and needs a lot of human input.
- Generative AI (such as ChatGPT) has limited autonomy since it requires human input to generate output. It does not initiate a set of actions independently.
- Fully autonomous systems work without intervention or supervision by a human. These AI systems are not currently used in the UK.
Wherever human intervention is required, such as for Generative AI systems, there is the possibility of human error that could result in losses for customers. Whilst this liability more squarely fits with professional indemnity, if the human error can be attributed to management oversights, a D&O claim could follow.
“Silent” AI cover
Whether AI risks are covered under a policy will depend on the wording. However, D&O wordings may not necessarily need to define 'AI risks' for losses arising from the use of AI to be covered. Here's why:
- Side ABC “Wrongful Act”: This is a key concept to cover which is typically broadly defined and will likely cover decisions relating to a policyholder's deployment and use of AI.
- Side A / B claim: “Investigation”: This is also broadly defined to include regulatory investigations, which would include investigations arising from the deployment or use of AI. For example, businesses within the scope of the EU AI Act's regulations may be exposed to an investigation or enforcement action by the AI Office. Such risk is likely already covered by a D&O wording.
- Side C claim "AI Washing": This is where a company exaggerates or misrepresents the capabilities of its AI system to secure investment. Whilst AI washing is a developing concept, claims by shareholders arising from alleged misstatements by officeholders will typically be covered under "Side C" of existing D&O policies.
Bringing it to life
The regulator begins formal proceedings against the company’s board, questioning whether directors failed to adequately oversee the AI’s deployment and governance.
We expect that most D&O wordings would likely cover this investigation as the policy typically responds to claims or investigations alleging wrongful acts by directors in their management capacity. The investigation into whether the board failed to oversee the AI system properly would be considered a management act or omission, falling within the scope of Side A or B coverage. We expect that the policy would therefore likely cover defence costs and, where appropriate and dependent on the exact policy wording, certain penalties or settlements arising from the investigation.
This scenario highlights the importance of reviewing D&O wording in light of fast-evolving regulatory frameworks - such as the EU AI Act or sector-specific guidance - to ensure directors remain protected as new AI risks and regulatory expectations emerge.
“Affirmative” AI Cover
Whilst it may provide helpful clarity, it's important to carefully define what's covered to avoid narrowing cover unintentionally. For example, the definition of “AI” will need to be broad enough to capture the existing AI systems listed above as well as any new ones that are developed, to avoid reducing the scope of cover available.
Stand-alone products are yet to be truly tested. In the absence of any exclusion in the policy, the same risks covered by AI policies may well be covered under existing D&O insurances.
AI Exclusions
Proactive policy wording
The risk landscape is constantly evolving, and where questions arise, underwriters and brokers will undoubtedly collaborate to find the best solution for policyholders.
It’s essential that brokers are proactive in reviewing policy wordings and guiding clients through what’s covered, particularly as we see the AI landscape develop.
