Extracted from Law360
The rise of artificial intelligence has fueled a new wave of securities class actions, with 34 cases filed since 2023, including 12 in the first half of this year alone. At the current pace, 2025 is on track to break the record for AI-related securities class action filings.[1]
As AI-related securities cases advance through the courts, early rulings on motions to dismiss are shaping this emerging legal landscape. We analyzed the nine motions to dismiss decisions issued in AI-related securities class actions filed between Jan. 1, 2023, and June 30, 2025, and identified some key trends for practitioners and corporate counsel managing AI-related litigation risks.
Despite early predictions that AI-related cases would be more likely to survive a motion to dismiss, the cases from this period were no more likely to survive than average. In all nine cases during this period, the courts granted a defendant's motion to dismiss in whole or in part.
From these rulings, three key trends stand out.
1. Scrutiny of Statements About AI's Capabilities
Courts continued to scrutinize representations about companies' AI capabilities, repeatedly emphasizing that plaintiffs must allege specific, concrete facts demonstrating that companies' AI did not have those purported capabilities to withstand a motion to dismiss.
For instance, in Lamontagne v. Tesla Inc., the U.S. District Court for the Northern District of California dismissed claims on Sept. 30, 2024, that Tesla exaggerated its self-driving technology's no-human intervention capabilities, because the plaintiffs failed to provide specific, contemporaneous facts showing the technology could not achieve zero-intervention drives as described.[2]
The court also declined to credit allegations based on a confidential witness's assertion that Tesla's autopilot department "absolutely" knew the statements were exaggerations, finding the allegations too vague "to show that [Elon] Musk's statements were false or misleading."
Likewise, the U.S. District Court for the Eastern District of Michigan's March 28 decision in In re: General Motors Co. Securities Litigation found the Cruise LLC subsidiary's statement describing self-driving cars as "safer than a human driver" inactionable, because the complaint did not allege facts allowing a meaningful comparison between the cars and human drivers.[3]
Although the complaint described how the cars "tend to drive 'erratically,'" it did not address whether human drivers also engage in similar behaviors.
Even in the context of statements about new, fast-paced technological developments like AI, courts still require detailed, comparative factual allegations to support claims that statements about a technology's capabilities are materially misleading.
2. Scrutiny of Claims About Who Is Doing the Work — AI or Humans
Courts are closely examining claims where companies allegedly represented human work as AI-driven.
For example, in In re: GigaCloud Technology Securities Litigation, the U.S. District Court for the Southern District of New York found on Jan. 27 that the plaintiff sufficiently pled that GigaCloud's statement that it used "complex AI software in [its] technology infrastructure" was false, since the company in fact "relied on manual computations" requiring "100 IT employees to maintain."[4] The court largely credited these allegations because of supporting statements from IT and supply chain employee confidential witnesses.
Additionally, in the General Motors case, the court found that a reasonable investor could plausibly interpret statements by Cruise's CEO that self-driving cars operated with "the human out of the loop" to mean the vehicles operated without any human input at all, not merely without a safety driver. The court declined to dismiss the claims, noting the defendants did not dispute that humans were involved in operating the vehicles.
In contrast, the Eastern District of Michigan granted the defendants' motion to dismiss other statements that the cars were "fully driverless" or "fully autonomous," concluding that a reasonable investor would interpret them to mean that self-driving cars could operate without a safety driver physically present while still allowing for some human involvement, such as remote operators.
Because GM and Cruise consistently clarified that "fully driverless" referred only to the absence of a safety driver in the vehicle and disclosed the use of remote operators, the court held that the plaintiffs failed to plead falsity.
These cases illustrate that courts are scrutinizing statements about AI performing work without human labor or intervention, but — as in other securities cases — plaintiffs must specifically allege facts showing those statements were false.
3. Issues Unrelated to AI Still Drive Many Outcomes
The decisions on motions to dismiss from this period also highlight that many AI-related cases, including multiple cases involving companies offering AI-related products or services, will continue to turn on issues that aren't AI-specific, such as claims about business prospects or scienter.
For instance, in In re: Vicor Securities Litigation, the Northern District of California considered claims that Vicor's statements about an agreement with a significant and existing customer for Vicor's AI-related service were misleading.[5] The court dismissed these claims on June 6 as forward-looking and expressed as expectations, not assurances of a concrete deal, and also found the complaint failed to plead scienter.
These cases illustrate how courts may resolve securities claims involving AI by applying established pleading standards rather than focusing on the unique nature of AI-related disclosures.
Takeaways
With this wave of litigation gaining momentum and courts providing initial guidance, companies should understand these distinctions to effectively manage risk and disclosure. Here are some practical steps.
Review and substantiate AI-related statements.
Ensure all public statements about AI capabilities, performance, and integration are accurate and substantiated. Avoid broad or unqualified comparisons between AI and human performance unless they can be supported by data.
Clarify the role of human labor.
Be clear about the extent of human involvement in processes described as AI-driven. Explicitly distinguish between tasks performed by AI, and those requiring human oversight, intervention or manual work. Misrepresenting human work as AI can create significant litigation risk.
Maintain established disclosure practices.
Continue applying established standards of accuracy, transparency and appropriate cautionary language when disclosing AI developments. Courts are still evaluating these disclosures under traditional securities law principles, so when in doubt, rely on those standards.
Monitor evolving case law.
Track new decisions in this rapidly developing area. A record number of AI-related cases are expected this year,[6] and nearly three-quarters of defense counsel expect the trend to continue, according to a survey conducted by Inigo Insurance.[7] Early rulings are shaping the standards for pleading and proving AI-related misstatements.
Conclusion
AI may be new, but the courts are applying the same exacting standards that have long governed securities litigation.
[1] Cornerstone Research, Securities Class Action Filings: 2025 Midyear Assessment (2025).
[2] Lamontagne v. Tesla, Inc. , No. 23-cv-00869 (N.D. Cal.).
[3] In re: Gen. Motors Co. Sec. Litig., No. 23-CV-13132 (E.D. Mich.).
[4] In re: GigaCloud Tech. Inc. Sec. Litig. , No. 23-cv-10645 (S.D.N.Y.).
[5] In re: Vicor Securities Litigation , No. 24-cv-04196 (N.D. Cal.).
[6] Cornerstone Research, Securities Class Action Filings: 2025 Midyear Assessment( 2025).
[7] Inigo Insurance, Inigo Defense Counsel Survey 2025 (2025).