Describing your AI to investors and customers: what regulators have acted on

Investors and customers decide based on what you tell them. If the company's AI does less than the deck or the website says, the difference can become a legal problem. US regulators acted on such differences in 2024, 2025 and 2026. The Securities and Exchange Commission (SEC) acts on statements made to investors. The Federal Trade Commission (FTC) acts on statements made to customers. This page sets out their cases about AI claims, from the agencies' own releases, and turns them into a check you can run on your own deck and site. It is part of the guide on building your AI product after you raise.

Several of these cases are charges or allegations, which the agencies must still prove. Where a case was settled, the company agreed to the order without the facts being decided by a court. Each paragraph below says which.

The SEC and statements to investors

Delphia and Global Predictions, March 2024

On March 18, 2024, the SEC "announced settled charges against two investment advisers, Delphia (USA) Inc. and Global Predictions Inc., for making false and misleading statements about their purported use of artificial intelligence (AI)" [1]. The release says "Delphia agreed to pay a civil penalty of $225,000, and Global Predictions agreed to pay a civil penalty of $175,000" [1], a total of $400,000. The firms settled without admitting or denying the findings. According to the SEC, Global Predictions "falsely claimed to be the 'first regulated AI financial advisor'" [1].

The SEC's chair at the time, Gary Gensler, named the practice: "Investment advisers should not mislead the public by saying they are using an AI model when they are not. Such AI washing hurts investors" [1]. AI washing means claiming more use of AI than a company has. The SEC's enforcement director, Gurbir S. Grewal, added that "public issuers making claims about their AI adoption must also remain vigilant about similar misstatements that may be material to individuals' investing decisions" [1]. Public issuers are companies that sell shares to the public.

Joonko, June 2024

On June 11, 2024, the SEC "charged Ilit Raz, CEO and founder of the now-shuttered artificial intelligence recruitment startup Joonko, with defrauding investors of at least $21 million" [2]. The SEC alleges that "Raz falsely told investors that Joonko had more than 100 customers, including Fortune 500 companies", and that Raz "lied to investors that Joonko had earned more than $1 million in revenue and was working with more than 100,000 active job candidates" [2]. These are allegations. A parallel criminal case was announced the same day, and this guide did not check the outcome of either.

Nate, April 2025

On April 9, 2025, the SEC "charged Albert Saniger, the founder and former CEO of Nate, Inc., a privately held technology startup, with fraudulently soliciting investments in Nate and raising over $42 million through the sale of Nate stock by making false and misleading statements about the company's use of artificial intelligence" [3]. The SEC alleges that Saniger "told investors in both Nate's Seed and Series A fundraising round that Nate's application used automated technology that relied on AI to complete purchases made through the app without human involvement", while "Nate relied in large part on contract employees to manually input orders placed by users on the app" [3].

Nate was a private startup, which shows the SEC acting on statements to private investors. The statements were also made in a seed and a Series A round, the same rounds this guide is about. The SEC listed the Nate case among its results for fiscal year 2025 in a release of April 7, 2026 [6]. These are allegations, and this guide did not check the outcome.

Presto Automation, January 2025

On January 14, 2025, the SEC issued an order against Presto Automation, which sold voice ordering for restaurant drive-thrus. The SEC found that "Presto failed to disclose that it used Supplier A's AI technology, rather than its own technology (which it did not begin to develop until early 2022), to power all Presto Voice units it commercially deployed during this time period" [4]. It also found that the company "misrepresented the capabilities of the product by claiming that it eliminated the need for human order taking", while "Presto hired, trained, and supervised human order takers located abroad (primarily in the Philippines and India), who processed the vast majority of drive-thru orders placed through Presto Voice" [4].

The order is a cease-and-desist order, which tells the company to stop. The Commission says it "considered Respondent's current financial condition" [4], and the order's sanctions section names no money penalty.

A unit for these cases

On February 20, 2025, the SEC announced its Cyber and Emerging Technologies Unit. Its listed priorities include "Fraud committed using emerging technologies, such as artificial intelligence and machine learning" [5].

The FTC and statements to customers

Operation AI Comply, September 2024

On September 25, 2024, the FTC announced "five law enforcement actions against operations that use AI hype or sell AI technology that can be used in deceptive and unfair ways" [7]. Its chair at the time, Lina M. Khan, said: "The FTC's enforcement actions make clear that there is no AI exemption from the laws on the books" [7].

DoNotPay, finalized February 2025

One of those actions was against DoNotPay, which advertised an AI lawyer service. The FTC's complaint alleges "that the company did not conduct testing to determine whether its AI chatbot's output was equal to the level of a human lawyer, and that the company itself did not hire or retain any attorneys" [7]. The final order, announced February 11, 2025, "requires DoNotPay to pay $193,000 in monetary relief" and "prohibits DoNotPay from advertising that its service performs like a real lawyer unless it has sufficient evidence to back it up" [8].

Workado, April and August 2025

On April 28, 2025, the FTC announced an order against Workado. It "settles allegations that Workado promoted its AI Content Detector as '98 percent' accurate in detecting whether text was written by AI or human. But independent testing showed the accuracy rate on general-purpose content was just 53 percent, according to the FTC's administrative complaint" [9]. The FTC added that "Each violation of such an order may result in a civil penalty of up to $53,088" [9], a figure stated on that date and adjusted over time.

The final order, approved on August 28, 2025, says Workado "Is prohibited from making any representations about the effectiveness of any AI content detection product unless it is not misleading, and the company has competent and reliable evidence to support the claim at the time it is made", and "Is required to retain evidence to support such efficacy claims" [10]. The Commission voted 3-0 to approve it [10].

Evolv Technologies, November 2024

On November 26, 2024, the FTC took action against Evolv Technologies "over allegations that the company made false claims about the extent to which its AI-powered security screening system can detect weapons and ignore harmless personal items, including in school settings" [11]. Samuel Levine, the FTC's consumer protection director named in the release, said claims about technology, including artificial intelligence, need to be supported by evidence [11].

Air AI, March 2026

On March 24, 2026, the FTC announced a proposed settlement with Air AI and its owners, who will be banned from marketing business opportunities. The order "includes a monetary judgment of $18 million, which will be largely suspended based on the company's and operators' inability to pay the full amount, requiring the operators of Air AI to pay $50,000 to the Commission for consumer relief" [12]. This case is about the earnings and refund claims an AI company made to small businesses.

The cases in one table

Case and date	Agency	The claim at issue	Status
Delphia and Global Predictions, March 2024	SEC	Use of AI that the firms did not have	Settled, $400,000 in total [1]
Joonko, June 2024	SEC	Customers, revenue and candidate numbers	Charges [2]
Evolv Technologies, November 2024	FTC	What its AI weapons scanner could detect	Action announced [11]
Presto Automation, January 2025	SEC	Whose AI ran the product; the claim that no human order takers were needed	Cease-and-desist order, no money penalty named [4]
DoNotPay, February 2025	FTC	AI service performs like a lawyer	Final order, $193,000 [8]
Nate, April 2025	SEC	AI completed purchases without humans	Charges [3]
Workado, April and August 2025	FTC	AI detector 98% accurate	Final order; must keep evidence [10]
Air AI, March 2026	FTC	Earnings and refund claims	Proposed order, $18 million judgment, $50,000 to pay [12]

What these claims had in common

The claims in the table fall into five types:

1. How much is automated. Nate and Presto described work as done by AI that, according to the SEC, people did.
2. Whose technology it is. Presto did not disclose that a supplier's technology ran its deployed units.
3. An accuracy figure. Workado's 98% against an independent test result of 53%, according to the FTC.
4. A comparison with a professional. DoNotPay's claim that its service performed like a lawyer, which, the FTC's complaint alleges, the company had not tested against a lawyer's work.
5. Business numbers. Joonko's alleged customer, revenue and candidate counts.

Each type has a check you can run before the claim goes into a deck or onto a page.

Type of claim	The question to ask before you make it	The evidence to keep
"Fully automated" or "no humans"	Does any person do any part of this work today, even sometimes?	A description of every step and who does it
"Our own AI" or "proprietary model"	Which company's model runs each part of the product?	The provider and model list, dated
A percentage for accuracy	Measured on how many cases, which cases, and when?	The eval results, the cases and the date
"As good as a doctor, lawyer or accountant"	Was it tested against that professional's work?	The test design and results
Customers, revenue, users	Do the numbers match the books and the data room?	The records behind each number

Reveneau's view

Reveneau's view is that every claim about the AI in a deck, on a website or in a sales email should be checked against the product on the day it is made, and the result kept with the date. An accuracy claim needs an eval suite, a set of written, repeatable tests of the product's answers on cases you chose in advance. The FTC's final order against Workado sets this standard for that company: competent and reliable evidence at the time the claim is made, kept on file [10]. Holding that evidence is good practice for every company, whether or not a regulator ever asks. The guides on why AI evals matter and on how to verify an AI claim with a test explain the method, and what investors expect before the next round covers keeping the deck and the records in agreement.

Where Reveneau fits

This page holds Reveneau to the same rule, so its own claims here are written as its own statements. Reveneau is an AI software development consultancy with offices in San Francisco and Vancouver. Reveneau says AI writes all of its code and every change must pass an eval suite written from the specification. Reveneau says Jev, TypeSafe AI's decision model, grades that suite, and that the suite runs ten times faster than with its previous grader. It says it uses AI instead of hiring more engineers and puts that saving into clients' prices. To plan an AI product and test what you will say about it, see AI development at Reveneau or contact Reveneau.