Sam Cornelius, lead management liability underwriter at rrelentess, on the risks that comes with artificial intelligence for business leaders.

Artificial Intelligence (AI) is increasingly used in the workplace across sectors, job functions and varying levels of seniority. But what happens when AI starts influencing company decision-making at the board level?

While AI brings efficiency to the boardroom, it also introduces risks, including data bias and reputational damage if “AI washing” triggers litigation or regulatory action. AI is prone to hallucinations and requires appropriate human oversight. Without it, directors risk leaving themselves dangerously exposed.

For the Directors & Officers (D&O) market, the key question is no longer whether companies are using AI, but how that use is governed, and who remains accountable when board-level decisions go wrong. Without robust oversight frameworks, individuals may face increased exposure to management liability claims.

AI use at the board level introduces new risks

A 2026 UK survey found that nearly three-quarters of C-suite tech executives (73%) uploaded confidential company data into AI tools, with tech directors recording a similar figure (74%). Additionally, Lloyds Banking Group is one of the first blue-chip companies to acknowledge that it is rolling out AI board advice.

As AI adoption accelerates, brokers and underwriters are increasingly considering what this means for D&O exposure.

Questions being asked across the market include:

• How does AI use affect policy coverage and portfolio performance?
• What AI-related exposures should underwriters assess during placement?
• How should brokers advise insureds on emerging governance expectations?

Directors remain bound by their duties under the Companies Act 2006 to act in the best interests of the company. That responsibility does not fall away simply because decisions are informed, or even generated, by AI.

New governance roles may be required

AI can be a beneficial resource for smaller or mid-sized organisations which may lack the full range of professional support or advice available in-house or on demand at large organisations. Potential board-level uses include:

• Data analysis and management reporting
• Legal document and contract review
• Board meeting minutes and administration
• Compliance and process reviews

However, these use cases are not without risk. AI is prone to making mistakes, which introduces greater requirements for effective oversight. This may lead to new corporate leadership positions focused on AI governance.

Brokers and underwriters will need to ensure that D&O policies can capture these individuals within the scope of cover and respond to subsequent actions brought against them.

Regulatory exposure

When AI is embedded to make or support significant decisions, it raises questions about accountability – who is responsible if AI makes a mistake? Will we see cases where directors are held liable for relying on incorrect information or insufficient due diligence?

The UK government’s decentralised approach to AI regulation does not make accountability clearer, leaving it up to sector-specific regulations. For instance, the Financial Conduct Authority (FCA) expects firms to adopt appropriate governance measures, including board oversight of the organisation’s AI use. Failure to do so could lead to breaches of the FCA’s Principles for Businesses and potential enforcement action, especially in cases where there is a genuine risk of consumer harm.

For brokers and underwriters, this means AI exposure cannot be assessed in isolation. The relevant regulatory landscape will vary significantly by insured and sector, making a nuanced understanding of applicable standards increasingly important.

AI Washing

Another emerging exposure is so-called “AI washing,” where companies exaggerate or misrepresent their AI capabilities, development or deployment.

This risk of class action arising from AI-washing is more developed in the U.S., with Innodata serving as an example, but it is also relevant to UK D&O markets. In the case of Innodata, the software-based data management company faced a securities class action lawsuit as the plaintiff claimed the company misrepresented the extent to which its products and services use AI.

In the UK, there are mechanisms to facilitate potential actions under the Financial Services and Markets Act 2010. Additionally, regulators can bring actions against companies and directors for making incorrect or misleading communications to the market.

For brokers, this reinforces the need to understand not only whether an insured uses AI internally, but also how AI forms part of its external proposition and marketing.

AI bias and discrimination

Large Language Models (LLMs) are computer programmes fed on existing data. If the model is based on existing biased data, then the output will be biased too. This means AI can be discriminatory.

This is particularly relevant in the hiring process. Three in 10 UK employers are using AI in hiring, according to research by StandOutCV. However, without the right protections and governance, directors can be exposed to employment and reputational claims if AI enforces biases, even unintentionally.

As a result, underwriters and brokers should consider AI use across the entire business practice to understand how these tools are used and how their output is checked and verified by human eyes.

As AI adoption continues, brokers have an increasingly important role in helping clients understand how AI governance may affect their D&O risk profile. Key underwriting decisions are likely to factor in how AI is being used across the business and at the board level, whether formal AI governance is in place, which individuals are responsible for AI oversight and how AI outputs are reviewed, challenged and verified by humans. Ultimately, AI may improve boardroom efficiency, but it does not transfer or diminish directors’ legal duties. As regulatory frameworks mature and litigation develops, the adequacy of AI governance will become increasingly important in assessing D&O exposure.