On January 14, 2025, the U.S. Securities and Exchange Commission (SEC) charged Presto Automation Inc. (Presto) with violations of the Securities Act of 1933 and the Securities Exchange Act of 1934 for misleading artificial intelligence (AI) claims made by Presto in its SEC filings and public statements.1 Without admitting or denying the SEC’s findings, Presto consented to a cease-and-desist order.
The proceeding centered on the practice of “AI washing,” where a company exaggerates or distorts its AI capabilities, creating a false impression of the company’s technological expertise and future prospects. The SEC has increased its focus on this matter and issued guidance reminding companies that AI disclosures remain subject to the basic principles of the securities laws and should only be made with reasonable bases, which should be disclosed to investors.2 In addition to this informal guidance and recent administrative proceedings, the SEC has used the comment letter process to request enhanced disclosure of AI claims in registration statements and periodic reports.
Background
Presto, a restaurant technology company, claimed to have developed an innovative AI product, Presto Voice, that used AI-assisted speech recognition to automate drive-through order-taking at quick-service restaurants. Despite the marketing of Presto Voice as a “cutting-edge solution,” the SEC found that Presto (i) misled investors by failing to disclose that the voice AI technology powering most of Presto Voice was actually developed, owned, and operated by a third party, and (ii) exaggerated its own in-house technology’s capabilities by downplaying the level of required human intervention in the order-taking process.
In its disclosures, Presto referred to Presto Voice as “Presto’s speech recognition technology,” “Presto’s Voice product,” “Presto’s [V]oice ordering solution,” and “our technology.” The SEC found that not disclosing Presto’s reliance on a third-party supplier was materially misleading because it led investors to believe the technology was developed and owned by Presto. The company also failed to inform investors about the extent of Presto’s reliance on a third party to power Presto Voice at locations of its largest customer.
Additionally, Presto claimed that its in-house technology eliminated human involvement in the order-taking process. However, the technology actually required substantial human intervention to operate. According to the SEC, Presto used terms like “automated order completion” and “non-intervention” rates in investor presentations to create the impression that Presto Voice could complete orders without human involvement. In reality, Presto employed human agents in off-site locations, such as the Philippines and India, to process orders. Presto eventually acknowledged that over 70 percent of orders processed through its in-house version of Presto Voice required human intervention, and 100 percent of orders at certain locations required human intervention, with an average non-intervention rate of only 85 percent.
The SEC found that Presto understood that its terminology was likely to mislead and confuse investors about the true capabilities of Presto Voice but failed to correct the misleading statements until it became aware of the SEC’s investigation.
Key Takeaways From Presto
1. Establish a reasonable basis for AI-related disclosures
Companies should establish and disclose a reasonable basis for AI-related claims in their regulatory filings with the SEC.3 When disclosing the status of their AI technology, they should draw a clear distinction between current capabilities and future plans to avoid misleading investors about the true progress and readiness of their AI offerings. In commenting on AI-related disclosures included in registration statements and periodic reports, the SEC has frequently asked companies to:
- Describe the scope of the current capabilities of AI technology
- Clearly distinguish current technological capabilities from future aspirational capabilities and product and service offerings
- Describe any material steps that will need to be taken to fully develop the technologies involved
As illustrated by the investigation of Presto, companies should fairly describe the stage of development and operational functionality of their AI products and services. For example, claims of “AI-driven” or “fully autonomous” systems should be accurate and distinguish between what is currently achievable and what is aspirational. Overstating AI capabilities — such as claiming that AI eliminates human involvement — without disclosure of its limitations could expose companies to regulatory scrutiny and legal liability.
2. Clearly define capabilities
Companies should define key AI-related terms, such as “generative AI” or “machine learning,” within the context of their business operations. The SEC recommends offering clear definitions to help investors understand the nature and scope of the AI technologies being discussed. SEC comments have included requests for clarification on whether “machine learning” and “AI technologies” have different definitions. To the extent that they have different meanings and function differently within a company’s platform, these differences should be explained.
3. Disclose reliance on third-party providers and related risks
Additionally, disclosures should clearly identify the involvement of third-party providers in supplying of AI technologies. SEC comments on this topic have requested clarification regarding whether companies utilize third-party products or internally developed AI processes, including whether companies intend to utilize open-source technology or license the use of such technology. To the extent companies intend to license existing or future technology, or plan to utilize proprietary and/or open-source technology, companies should add relevant risk disclosures to address any related risks.
Companies should also assess the impact on operations if a third party terminates the relationship and whether risk disclosures would be necessary. Additionally, companies should evaluate whether any licensing agreements governing the use and scope of AI technology should be described and/or filed.
In Conclusion
If a company deems the use of AI in its operations, products, or services material enough to warrant disclosure in public filings or statements, such disclosures must carefully adhere to the fundamental principles of federal securities laws. AI-related disclosures should provide investors with material information and be grounded in reasonable bases. As demonstrated by Presto, the SEC will be closely examining such disclosure for accuracy and completeness. Given the growing adoption of AI across industries worldwide, public companies are urged to approach these disclosures with heightened attention and diligence.
1 See U.S. Securities and Exchange Commission, In the Matter of Presto Automation, Inc., Release No. 33-11352, Jan. 14, 2025, available at https://www.sec.gov/files/litigation/admin/2025/33-11352.pdf.
2 See U.S. Securities and Exchange Commission, “Office Hours with Gary Gensler: AI Washing,” Sept. 4, 2024, available at https://www.sec.gov/newsroom/speeches-statements/gensler-office-hours-ai-washing-090424; see also U.S. Securities and Exchange Commission, “Statement from Erik Gerding: State Disclosure Review,” June 24, 2024, available at https://www.sec.gov/newsroom/speeches-statements/gerding-statement-state-disclosure-review-062424.
3 Establishing and documenting the reasonable bases for AI-related disclosures should be incorporated into a company’s existing disclosure controls. In Presto, the SEC found that Presto did not have an established process to ensure that the information required to be disclosed was recorded, processed, summarized, and reported, nor was it communicated to Presto’s management. The result was that no one at Presto was formally responsible for ensuring that Presto’s disclosure was accurate.