By Christine Polek (April 2, 2026)

Since the launch of ChatGPT in 2022, references to artificial intelligence in public company disclosures have surged.[1] By 2025, roughly 90% of large U.S. companies referenced AI in their annual reports,[2] and more than 40% of S&P 500 companies discussed AI during earnings calls.[3]

Across industries, companies now routinely frame their growth strategies around AI integration and AI-enhanced efficiencies.

Some companies are fundamentally repositioning their business for an AI-driven economy, while others are layering AI into their existing business operations.

For example, public safety technology company Axon Enterprise Inc. has begun repositioning its business from a hardware-focused model toward a software-and data-driven platform built around AI-enabled services, telling investors in February that "we are positioned to be a winner in this AI-driven environment and we intend to lap the field".[4]

Similarly, document-management service company Adobe Inc. has emphasized a shift toward an AI-enabled platform, stating in September 2023 that "we're bringing generative AI to life across our portfolio of apps and services to deliver magic and productivity gains."[5]

By contrast, other companies emphasize AI as an enhancement to existing offerings rather than a fundamental transformation. Home improvement retailer Lowe's Cos. Inc. told investors in February 2025 that is "leaning into emerging technology to enhance the customer experience, saving them both time and money,"[6] while traditional service firm Accenture PLC told investors in September 2025 that it had "worked on more than 6,000 advanced AI projects just this year" and "we are now delivering differently for our clients."[7]

Research suggests that, overall, the tone of these disclosures is systematically optimistic. In a November 2024 working paper from the National Bureau of Economic Research, sentiment analysis of transcripts of earnings calls showed that AI-related commentary skewed strongly positive.[8] Words such as "efficiently," "tremendous" and "advantageous" appear far more frequently than negative descriptors. Since 2017, when AI references began appearing with regularity in corporate disclosures, more than 30% of AI mentions in earnings calls have carried positive sentiment, compared with roughly 10% that were negative.[9]

Does this consistent optimism pose legal risks? Under the traditional doctrine of the U.S. Securities and Exchange Commission's Rule 10b-5, courts often throw out securities fraud claims when the challenged statements at issue are simply too vague or puffy to be actionable. Broad, feel-good assertions that cannot readily be proven true or false are usually considered immaterial as a matter of law.

For example, when Align Technology Inc., the maker of Invisalign clear aligners, described China as a "great growth market" and characterized its performance there as "tremendous," the U.S. Court of Appeals for the Ninth Circuit's 2022 opinion in Macomb County Employees' Retirement System v. Align, affirming the U.S. District Court for the Northern District of California's previous dismissal, treated those comments as ordinary corporate optimism rather than specific, factual claims capable of misleading a reasonable investor.[10]

Yet with respect to optimistic AI disclosures, empirical evidence suggests that investors may in fact treat such statements as economically meaningful.

In late February, AI-related developments drove sharp swings in major indices and individual equities. Salesforce.com Inc., one of the world's largest technology companies, reported strong fourth-quarter and full-year results, but its stock fell in after-hours trading amid broader market concerns about AI-driven disruption.[11]

Like Salesforce, other software-as-a-service, or SaaS, companies were, per Business Insider on Feb. 23, "pummeled by renewed fears of AI disruption,"[12] losing billions in market value in what commentators dubbed a "SaaSpocalypse." The following day, however, the market recovered a substantial portion of those losses after getting what the Associated Press described as "a reminder that the artificial-intelligence boom may also have an upside."[13]

Notably, these movements were driven primarily by forward-looking concerns about competitive positioning in an AI-enabled SaaS landscape, rather than contemporaneous changes in fundamentals.

Event study research examining AI-related disclosures in Form 10-K filings provides additional context. A September 2025 paper in the International Review of Financial Analysis showed that firms issuing initial, substantive AI disclosures experience statistically significant positive abnormal returns, suggesting that investors reward credible signals of AI adoption and strategic positioning.[14] These findings suggest that investors are not merely reacting to the presence of AI-related language, but are actively assessing whether the statements are economically meaningful.

At the same time, the rise in AI-related securities litigation appears to reflect the ambiguity often embedded in corporate statements about AI. In 2025, approximately 16 AI-related securities class actions were filed, representing about 8% of all federal securities class actions that year.[15] Many of these cases center on allegations that companies overstated their AI capabilities or the financial impact of AI initiatives, a practice sometimes described as AI washing.

These cases test the boundary between nonactionable puffery and materially misleading statements. Describing a company as "AI-driven" or "AI-enabled" may constitute the kind of generalized optimism courts routinely dismiss as immaterial. However, when a company ties those characterizations to specific representations, such as customer adoption or projected AI-related revenue, the analysis shifts. What might initially sound like marketing language can become a concrete, verifiable claim.

The Northern District of California's August 2025 decision in Dolly v. GitLab Inc. illustrates the difficulty of drawing this line. Investors alleged that GitLab overhyped the AI capabilities of its software and misled the market about customer demand, citing statements such as "we already have 14 AI features available to our customers."[16]

After the stock price fell, the plaintiffs alleged that those representations had artificially inflated market expectations. The court dismissed the complaint, concluding that the statements were either forward-looking or nonactionable puffery, and that the plaintiff failed to show falsity or intent.[17]

The ruling is consistent with the economic intuition that markets tend to discount generalized or promotional statements and instead prize verifiable, economically meaningful information. At the same time, market experience and academic research suggest that disclosures about AI features, integration or competitive positioning can influence expectations about future cash flows, even when framed in high-level terms.

So, when does an AI-related statement cross the line from optimistic framing to actionable misrepresentation? As AI becomes increasingly central to valuation across industries, the materiality inquiry may increasingly focus less on tone and more on economic substance. Distinguishing technological transformation from strategic spin requires careful, data-driven analysis of whether and how AI initiatives translate into sustainable competitive advantage.

In the AI context, that inquiry may require technical understanding of the claimed AI capabilities. Ultimately, the key economic question is whether the statement provides information that a rational investor would use to update a valuation model, which may depend, in part, on a rigorous assessment of the technology itself.

This piece was originally published by Law360: https://www.law360.co.uk/articles/2457143/when-ai-puffery-becomes-actionable-securities-fraud

Christine Polek is a senior principal at Keystone AI.

The opinions expressed are those of the author(s) and do not necessarily reflect the views of their employer, its clients, or Portfolio Media Inc., or any of its or their respective affiliates. This article is for general information purposes and is not intended to be and should not be taken as legal advice.

[1] Michele Ca' Zorzi, Gianluigi Lopardo, Ana-Simona Manu, What Corporate Earnings Calls Reveal About the AI Stock Rally, VoxEU (CEPR), Sept. 18, 2025, https://cepr.org/voxeu/columns/what-corporate-earnings-calls-reveal-about-ai-stock-rally.

[2] Center for Audit Quality, S&P 500 and AI Reporting (2025), https://www.thecaq.org/sp-500-and-ai-reporting.

[3] FactSet, More Than 40% of S&P 500 Companies Cited AI on Earnings Calls for Q2 (Aug. 2024), https://insight.factset.com/more-than-40-of-sp-500-companies-cited-ai-on-earnings-calls-for-q2.

[4] Axon Enterprise Inc., Q4 2025 Earnings Call Transcript (Feb. 24, 2026) (statement of Joshua Isner), available at https://filecache.investorroom.com/mr5ir_axon/592/Axon_Enterprise_Inc_Q4_2025_Earnings_Call_Transcript.pdf.

[5] https://www.adobe.com/cc-shared/assets/investor-relations/pdfs/q3-fy2023-earnings-transcript.pdf.

[6] Lowe's Cos. Inc., Q4 2024 Earnings Call Transcript (Feb. 26, 2025) (statement of Marvin Ellison, CEO), available at https://corporate.lowes.com/sites/lowes-corp/files/2025-02/low-usq-transcript-2025-02-26.pdf.

[7] Accenture PLC R (CSA.HA) Q4 FY2025 earnings call transcript Powered by Quartr Sept. 25, 2025 https://investor.accenture.com/~/media/Files/A/Accenture-IR-V3/quarterly-earnings/2025/q4-fy-25/accenture-fourth-quarter-fiscal-2025-conference-call-transcript.pdf.

[8] Hassan, Tarek Alexander, et al. Economic surveillance using corporate text. No. w33158. National Bureau of Economic Research, 2024, available at https://www.nber.org/system/files/working_papers/w33158/w33158.pdf.

[9] Hassan, Tarek Alexander, et al. Economic surveillance using corporate text. No. w33158. National Bureau of Economic Research, 2024, available at https://www.nber.org/system/files/working_papers/w33158/w33158.pdf.

[10] Macomb County Employees' Retirement System v. Align Technology Inc., No. 21-15823 (9th Cir. July 7, 2022), available at https://cdn.ca9.uscourts.gov/datastore/opinions/2022/07/07/21-15823.pdf.

[11] Salesforce's Stock Falls as Mixed Earnings Forecast Fails to Dispel AI Gloom, MarketWatch (Feb. 25, 2026), https://www.morningstar.com/news/marketwatch/20260225444/salesforces-stock-falls-as-mixed-earnings-forecast-fails-to-dispel-ai-gloom.

[12] Jennifer Sor, Software Stocks Are Tanking the Market Again as AI and Tariff Uncertainty Spook Traders, Business Insider (Feb. 23, 2026), https://www.businessinsider.com/stock-market-today-selloff-tariffs-trump-software-ai-2026-2.

[13] U.S. Stocks Rise After Getting a Reminder of AI's Potential Upsides, AP News (Feb. 24, 2026), https://www.apnews.com/article/stocks-markets-trump-ai-iran-209a20c307b116ef8104c7166c8d2c9c.

[14] Anup Basnet, Maxim Elias, Galla Salganik-Shoshan, Thomas Walker, Yunfei Zhao, Analyzing the market's reaction to AI narratives in corporate filings, International Review of Financial Analysis, Volume 105, September 2025, https://www.sciencedirect.com/science/article/pii/S105752192500465X.

[15] Cornerstone Research, Securities Class Action Filings - 2025 Year in Review (2026), https://www.cornerstone.com/wp-content/uploads/2026/01/Securities-Class-Action-Filings-2025-Year-in-Review.pdf, and Secretariat, Trends in AI-Related Securities Class Actions Through 2025 (2025), https://secretariat-intl.com/insights/trends-in-ai-related-securities-class-actions-through-2025/.

[16] Dolly v. GitLab Inc., No. 5:24-cv-06244-EKL (N.D. Cal. Aug. 14, 2025), available at https://docs.justia.com/cases/federal/district-courts/california/candce/5:2024cv06244/435964/72.

[17] Dolly v. GitLab Inc., No. 5:24-cv-06244-EKL (N.D. Cal. Aug. 14, 2025), available at https://docs.justia.com/cases/federal/district-courts/california/candce/5:2024cv06244/435964/72.