Every few years, a corporate fraud case shows how far a company’s claims about its technology can drift from reality, and how long that gap can go unnoticed. The questions afterward are always the same. But where were the internal controls? The data was there if anyone wanted to look at it. Where were the compliance officers? Where were the board members? When concerns were raised, did anyone check? Too often, everyone wanted to believe the lie.

But what if your company is lying, and you don’t know it? Worse, even if you suspected something was amiss, you can barely confirm it. We are talking about overselling a company’s use of artificial intelligence (AI). Overselling its accuracy (“Through our use of AI we can guarantee outcomes of 89% or better”). Overselling how much your company uses it (“We have adopted and utilize AI in all of our clinical phases.”). Overselling to the public. Overselling to investors. What may first seem like mere puffery can quickly become an actionable fraud—as history has made abundantly clear.

Sponsors deploying AI in clinical trials are creating legal exposure unrelated to whether the AI “works.” Securities class actions target overstated technology capabilities; derivative suits challenge inadequate vendor oversight. AI provides new fact patterns for existing legal theories.

The danger lies in the widening gap between AI adoption speed and legal preparedness. Companies integrate AI into trial-critical functions while disclosure practices and board oversight remain calibrated for traditional software. The result: inadequately disclosed material AI dependencies and unrecognized governance risks.

This article examines shareholder litigation exposure when AI capabilities are overstated, vendor oversight fails, or AI-related problems impact clinical programs and stock prices.

The Emerging Shareholder Litigation Risk

There has been a marked rise in AI-related securities litigation, driven by “AI-washing,” which is the practice of exaggerating or misrepresenting a company’s use of AI to attract investors. While this trend has been most visible in the technology sector, the same risk profile is rapidly emerging for life sciences companies integrating AI into clinical development.

For sponsors using AI in clinical trials, litigation risk does not turn on whether the technology ultimately performs as intended. Rather, exposure arises when public statements about AI capabilities, use cases, or expected benefits outpace what the company can substantiate, or when material risks associated with AI deployment are not adequately disclosed. These risks may include technological limitations, reliance on third-party vendors, evolving regulatory expectations, and downstream impacts on trial integrity or outcomes. When such information is omitted or mischaracterized, investors may pursue securities class actions tied to alleged misstatements or omissions in company disclosures.

Securities Class Actions

The law is well established in securities-related lawsuits. The Securities Exchange Act prohibits the use of any “manipulative or deceptive device” in connection with the purchase or sale of securities,[1] including making untrue statements of material fact or omitting information necessary to render statements not misleading (known in the industry as “Rule 10b-5”).[2] Rule 10b-5 claims require plaintiffs to plead, among other elements, (i) a material misrepresentation or omission, (ii) scienter, (iii) reliance, (iv) economic loss, and (v) loss causation.[3]

The first and probably most important element, “materiality,” turns on whether a reasonable investor would consider the AI-related statement important in making an investment decision. Courts have emphasized that actionable misstatements typically involve specific representations on which investors can reasonably rely as reflecting concrete facts or outcomes. This distinction is already playing out in AI cases: generalized statements about innovation may be treated as nonactionable “puffery,” while concrete claims about performance, efficiency, or revenue impact are more likely to be deemed material if inaccurate.[4]

To understand how these theories play out in practice, consider a hypothetical clinical-stage sponsor that tells investors its proprietary AI platform is “fully deployed” to identify and enroll eligible patients across its pivotal program. Internally, the tool is still in a limited pilot at a handful of sites, and enrollment is running behind schedule. If the company later announces a significant delay to its pivotal readout and the stock drops sharply, investors would have a basis to allege that the company’s public statements created a materially false impression of the platform’s operational status. As noted above, plaintiffs in recent AI-related securities cases have advanced this same theory. The takeaway for clinical trial sponsors is direct: exaggerated statements about AI capabilities, and the concealment of known limitations, can give rise to securities suits for violations of Sections 10(b) and 20(a) of the Securities Exchange Act.

To illustrate where courts draw the line between nonactionable AI hype and actionable misrepresentation, consider a hypothetical sponsor whose executives promote its AI-driven trial design and site selection tools on earnings calls and in investor presentations. Courts examining AI-related securities claims have distinguished between two categories of statements. First, vague and promotional language—statements that the company is “excited about the transformative potential of AI,” or that AI is “changing how we think about drug development”—is likely to be dismissed as nonactionable puffery. Puffery refers to generalized, optimistic statements that “both on their own terms and in context, lack[] a standard against which a reasonable investor could expect them to be pegged.”[5] By contrast, more specific assertions that the company’s AI model “cuts screen-failure rates by 40% compared with traditional site selection,” or that it “reliably predicts which sites will meet enrollment targets,” can be treated as potentially actionable because investors can reasonably test such claims against actual performance. For clinical trial sponsors, the lesson is clear: generalized enthusiasm about AI may be protected, but specific claims about superiority, adaptability, or measurable performance create securities litigation risk if they cannot be substantiated.

Shareholder Derivative Suits

In addition to securities class actions, AI adoption in clinical trials also creates potential exposure through shareholder derivative litigation. These claims typically allege breaches of fiduciary duties by directors and officers, particularly under duty-of-oversight theories.[6] Plaintiffs would have to show either “(1) that the directors knew or (2) should have known that violations of law were occurring and, in either event, (3) that the directors took no steps in a good faith effort to prevent or remedy that situation, and (4) that such failure proximately resulted in the losses complained of….”[7] In the AI context, such claims may arise where directors or officers:

1. Rely on AI to make a business decision that harms the corporation/company, leading to allegations of unreasonable reliance without further investigation into reliability and accuracy of the AI system utilized;
2. Fail to use commonly accepted AI systems to assist in the decision process, which results in inadequately informed decision making; or
3. Authorize an expensive but ineffective strategy implementing AI that results in significant harm to a company’s health and tarnishes its reputation.[8]

Overall, the litigation risk associated with AI in clinical trials is less about technological failure and more about governance, disclosure, and oversight. As AI becomes increasingly embedded in trial design, execution, and analysis, companies should ensure that their public statements, internal controls, and board-level oversight mechanisms evolve accordingly, or risk becoming early targets in this emerging wave of shareholder litigation.

Red Flags: What Makes Companies Vulnerable

• Public statements that create measurable AI performance expectations, particularly where companies describe AI as improving trial execution, data quality, efficiency, or outcomes in ways that investors can later test against actual performance. As the examples above illustrate, the risk increases when statements move beyond general optimism and imply concrete, verifiable superiority or operational capability.

• Board oversight gaps, including limited visibility into how AI is being used in clinical development, no regular reporting on AI-related risks or vendor dependencies, and no clear escalation path when AI issues could affect trial timelines, data reliability, or public disclosures. These are the kinds of governance weaknesses that create derivative-suit exposure under duty-of-oversight theories.

• Disclosure practices that emphasize AI upside while omitting known constraints, such as reliance on third-party tools, infrastructure limitations, or the possibility that AI-related problems could disrupt clinical programs or affect financial performance. As the examples in this article illustrate, companies become more vulnerable when market-facing statements understate the operational or governance risks that later materialize and move the stock price.

• Inadequate vendor contract oversight practices that are silent on AI, particularly where standard agreements rely on generic indemnification, warranty, and limitation-of-liability terms drafted for traditional software and never adapted to address AI-specific risks. In the clinical trial context, that gap can leave sponsors without meaningful recourse when a vendor’s AI tool misrepresents its capabilities, produces erroneous outputs, mishandles patient or trial data, or contributes to disclosure or trial-integrity failures that later surface in shareholder litigation.[9]

Conclusion

These shareholder litigation risks arise from gaps between claimed and substantiated AI capabilities and require actual board engagement and oversight.

The most vulnerable sponsors treat AI as purely operational. When AI becomes integral to clinical development, it triggers disclosure obligations and oversight duties.

Protection requires recognizing AI as a governance issue. Boards need visibility into AI tools, validation, and dependencies. Disclosure teams need materiality frameworks. Vendor oversight must address risks that generic indemnification clauses don’t cover.

These are known legal frameworks applied to new facts. Most sponsors haven’t recognized that their AI decisions are creating litigation traps for the unwary. Companies proactively addressing these gaps will avoid becoming test cases.

---

[1] 15 U.S.C. § 78j(b).

[2] 17 CFR § 240.10b-5(b).

[3] In re Philip Morris Int’l Inc. Sec. Litig., 89 F.4th 408, 417 (2d Cir. 2023) [internal citations omitted].

[4] See “Investors increasingly claim that AI hype is securities fraud,” Virginia Milstead, et al., Reuters Legal News, Aug. 7, 2025, https://www.skadden.com/-/media/files/publications/2025/08/investors_increasingly_claim_that_ai_hype_is_securities_fraud.pdf.

[5] City of Monroe Emps. Ret. Sys. v. Bridgestone Corp., 399 F.3d 651, 671 (6th Cir. 2005).

[6] See William E. Knepper & Dan A. Bailey, Liability of Corporate Officers and Directors § 1.19 (8th ed. 2025).

[7] In re Caremark Int’l Inc. Derivative Litig., 698 A.2d 959, 971 (Del. Ch. 1996).

[8] See William E. Knepper & Dan A. Bailey, Liability of Corporate Officers and Directors § 1.19 (8th ed. 2025).

[9] See “AI in the Fine Print: Why Your Contracts May Not Be Ready,” Kris Kappel and Maggie Mannebach, Healthcare Law Insights (May 2026), https://www.healthcarelawinsights.com/2026/05/ai-in-the-fine-print-why-your-contracts-may-not-be-ready/ (noting that most vendor agreements are silent on AI, general contract language was not drafted with AI in mind, and organizations should negotiate AI-specific warranties and indemnification covering privacy violations, erroneous outputs, and intellectual property infringement).