As artificial intelligence (AI) investment accelerates at a breakneck pace, questions are mounting about whether the boom has crossed into bubble territory, and what that could mean for corporate boards if expectations fall short.
According to an April 2025 report from the United Nations Conference on Trade and Development (UNCTAD), the global AI market could reach as much as $4.8 trillion by 2033, a 25-fold increase in just a decade. That growth projection has fueled massive spending on AI talent, data centers, and automation tools across industries.
For directors and officers, the concern is not whether AI will be transformative, but whether governance, disclosures, and capital allocation decisions made during this high-growth cycle could later be second-guessed by shareholders.
That risk is increasingly front of mind for D&O insurers, with some warning that AI may soon eclipse COVID-19 and crypto as a leading driver of directors and officers (D&O) litigation.
“Among event-driven federal securities class actions, COVID-19 and crypto have been leading causes over the past four years,” said Eric Wedin (pictured below), head of financial lines – North America at Allianz Commercial.
“However, AI is poised to surpass them and has the potential to drive a significant increase in directors’ and officers’ liability in the future.”
Comments from market leaders have fuelled speculation of a potential tech bubble. Last August, OpenAI CEO Sam Altman openly questioned whether investors were getting “overexcited” over AI. In a recent interview with BBC, Google boss Sundar Pichai warned there are “elements of irrationality” in the AI boom, noting that no company would be immune if the bubble were to burst.
Wedin said investors are closely watching how companies deploy AI, how quickly they do it, and what returns they promise. “If investments fail to perform as expected, or if companies fall behind their peers, shareholders may seek to hold directors accountable,” he said.
That accountability often comes in the form of securities class actions alleging misleading disclosures, overstated growth projections, or failures of oversight. These types of claims have historically targeted boards during periods of rapid technological change.
Nora Hattauer (pictured below), head of financial lines for Zurich North America, said the sheer scale of AI-related capital commitments, particularly around data center buildouts, is unprecedented, even for large technology firms with deep balance sheets.
“Any time capital commitments ramp up this quickly, it’s critical for boards to approach them with a very clear governance framework,” Hattauer said, adding that carriers are asking more pointed questions about accounting treatments, balance-sheet impacts, and liquidity implications related to AI and technology spending.
One of the biggest legal flashpoints is disclosure. Missed earnings guidance and misleading statements about future performance remain the leading causes of securities class action claims against directors and officers, and AI-driven projections are no exception.
Wedin noted a significant uptick in AI-related securities claim filings in the US during the first half of 2025. Citing a NERA report, he pointed out that 13 AI-related securities suits were filed in the first six months of 2025, compared with 16 for all of 2024. Many of these cases allege that companies overstated the benefits of AI or understated the risks to profitability.
“Even when a company includes safe harbor language, shareholders and plaintiffs’ attorneys will scrutinize decisions made by directors and officers if financial results don’t meet expectations,” Hattauer said.
Compounding the risk is a rapidly evolving regulatory environment. In December 2025, the Investor Advisory Committee of the US Securities and Exchange Commission released a draft recommendation urging the SEC to adopt AI-specific disclosure rules. Globally, regulators are taking divergent (and sometimes overlapping) approaches.
“With different, and sometimes overlapping, regulatory approaches used in many jurisdictions, it is a major challenge for D&Os to navigate these, especially for management of multinational companies,” said Wedin.
AI adoption also presents boards with a conundrum: move too fast, and risk allegations of mismanagement; move too slowly, and risk claims that directors failed to seize strategic opportunities.
A key vulnerability arises when companies rush AI tools into production without adequate testing. This includes reviewing training data for bias or intellectual property issues and validating outputs to avoid errors or so-called “hallucinations.”
“When adequate expertise is not present internally, external consultants should be engaged,” Wedin added. “Rushing AI tools into production without a well-thought-out process might leave D&Os vulnerable to allegations of mismanagement.”
At the same time, being perceived as overly cautious carries its own risks. “Being seen as too slow to implement an AI strategy or too conservative in its use could also represent a litigation exposure,” he said.
If AI-driven investments underperform, documentation may be one of the strongest defenses directors have against shareholder and regulatory scrutiny, according to Hattauer and Wedin.
“One best practice is documenting the ‘why’ behind major AI or data infrastructure investments,” she said. “Boards can put themselves in a much stronger position by showing they leaned on qualified experts – financial, technical, and legal – and that this input is reflected in their meeting minutes.”
“AI use must be subject to the same level of control as any other major strategic decision and public statement a company makes,” Wedin said. “Properly documenting each step of these control processes will go a long way to establishing that sound business judgment was used.”
Wedin also emphasized the importance of comprehensive reporting from management, robust governance frameworks, formal risk management processes, and ongoing education for directors.
For Hattauer, good governance in this environment means boards should be pressure-testing returns and preparing for unexpected or negative outcomes.
“These (AI) projects often span many years, and assumptions can evolve,” she said. “Directors should build in a process for revisiting the business case periodically and understanding how the company would respond if market conditions shift."
Underwriters are also closely watching how AI infrastructure is financed. Hattauer said Zurich is increasingly seeing the use of alternative financing structures, including special-purpose vehicles and other off-balance-sheet arrangements, to support massive data center projects.
“These structures aren’t necessarily inappropriate, but they warrant healthy caution,” she said. “D&O underwriters are skeptical and want to understand why a particular structure is being used and how risks are being managed.”
Liquidity is a particular concern for smaller or mid-sized companies eager to compete in the AI race. Boards must realistically assess whether their balance sheets can support the scale and timing of AI ambitions, especially if revenue growth lags expectations.
Separately, governance does not stop at infrastructure. Boards are also expected to oversee how AI is used operationally, including in hiring, underwriting, or customer interactions, where emerging regulation and liability risks (such as bias or model errors) are gaining attention, according to Hattauer.
For insurance brokers, it's critical to know that underwriters are looking for evidence of disciplined decision-making from boards, reliance on qualified experts, thoughtful disclosures, and a clear understanding of financing and liquidity risks.
“Those fundamentals can help mitigate D&O risk for companies and directors,” said Hattauer.
