Introduction

This alert highlights recent artificial intelligence (AI)-related enforcement actions that the Securities and Exchange Commission (SEC) has brought against investment advisers under the SEC's "Marketing Rule."¹ The enforcement actions generally allege that the investment advisers inaccurately marketed how they used AI and other algorithms and data in the investment process. Given the public interest in AI and similar technology, regulatory enforcement in this space is not surprising. This alert also highlights Marketing Rule enforcement actions related to firms’ use of hypothetical performance metrics in their advertising.

Takeaway

As always, investment advisers must ensure consistency between their marketing and their actual business practices. The SEC is focusing on buzzworthy topics like AI, algorithmic investing, and big data. Differences between what advisers claim they are doing and what they actually do can create liability under the Marketing Rule.

Marketing Rule Background

The SEC heavily regulates advertising by investment advisers. An investment adviser traditionally includes RIAs and private fund managers for long/short equity, private equity, and venture funds. Increasingly, investment advisers also include quantitative and data-driven hedge funds, crypto hedge funds, trading and other digital advisory apps, as well as certain non-discretionary research and analytics platforms. These technology-oriented firms tend to incorporate AI and other code into their investment processes more often. To the extent they advertise their advisory services, the firms are subject to the Marketing Rule.

Although the SEC's principles of regulating investment adviser advertising date back to the 1960s or even earlier, the Marketing Rule was revamped in 2020. Since then, the Marketing Rule has been an area of focus for the SEC. Since its adoption, there have been at least nine (depending on how you keep score) meaningful regulatory developments:

Notably, the timeline of these developments coincides with the recent rise of AI in the public sphere.

Below, we dive into two recent SEC enforcement actions that offer an initial look at the SEC's attitude on AI-related advertising by investment advisers.

Enforcement Actions: "AI Washing"

On March 18, 2024, the SEC announced settled enforcement actions against fintech companies Delphia (USA) Inc. and Global Predictions, Inc. involving alleged "AI Washing." According to the SEC, these companies marketed that they used AI, data, and other inputs in certain ways for their investors. The fintechs' actual practices, however, differed from what they advertised.

Delphia²

Delphia managed private investment funds and traded for more than 29,000 individual retail accounts. Delphia incorporated data and software into its trading and other investment processes, including using algorithms to manage portfolios based upon clients' differing investment objectives and risk profiles. Delphia intended to use AI and machine learning to collect data from its clients, such as information gathered from social media and banking and credit card transactions, and to feed the data into its algorithms. Delphia, however, never succeeded in using AI and machine learning to analyze client data. Yet, Delphia had already advertised on its website and in a press release that it did, in addition to making the claim in regulatory filings (Form ADV). For example, Delphia problematically claimed that:

• It used client data in "a predictive algorithmic model" in selecting stocks, ETFs and options; and 
   
• It used machine learning to analyze the collective data shared by its members to make intelligent investment decisions.

As mentioned, Delphia never used AI or machine learning to analyze client data. The SEC alleged in a settled enforcement action that Delphia violated the Marketing Rule, among other things. Delphia agreed to pay $225,000 in civil monetary penalties and to a cease-and-desist order and censure. 

Global Predictions³

Global Predictions provides investment tools and insights for retail clients. In particular, Global Predictions offers investment advisory services through PortfolioPilot, an online platform that uses algorithms to make investment allocation recommendations. According to the SEC, Global Predictions advertised on its website, social media, and elsewhere that:

• Global Predictions' technology incorporated "[e]xpert AI-driven forecasts"; and
   
• Global Predictions was the "first regulated AI financial advisor."

Unfortunately, according to the SEC, Global Predictions did not use AI as advertised. When pressed by the SEC, Global Predictions could not produce documents to substantiate its advertised use of AI. This violated a core tenet of the Marketing Rule, which requires advisers to have a reasonable basis for believing that they can substantiate material statements of facts in their advertisements. Additionally, the SEC alleged that Global Predictions inaccurately represented model performance, including by providing inaccurate depictions of the relevant time horizon and benchmarks used. We discuss performance advertising further below. Global Predictions settled charges that it violated the Marketing Rule, among other things, and agreed to a cease-and-desist order, a censure, and a $175,000 civil monetary penalty.

Takeaway

The Delphia and Global Predictions actions demonstrate that the SEC is focused on "AI washing" and overall Marketing Rule compliance. On a closer look, these settled enforcement actions, although cast as about AI, really were garden variety inaccurate disclosure cases. 

For example, in Delphia, the adviser’s main violation was rooted in inaccuracies about how it used data in research and trading-related processes. Delphia advertised that it used data in ways—by incorporating certain client specific data into its algorithms—that it simply did not. 

When advertising about AI, machine learning, other software and coding techniques, and data usage, advisers must accurately represent their actual business practices. "If you claim to use AI in your investment processes, you need to ensure that your representations are not false or misleading."⁴ Again, even though these topics are complex, preventing violations is not. Standard compliance processes, such as pre-approval of marketing pieces or using templates, can prevent regulatory headaches even for the most complex firms and funds. Finally, although this alert is about marketing, advisers should continue to educate themselves about risks associated with AI and other more complex code. Firms create code and can face liability for what it does.

Enforcement Actions: Hypothetical and Model Performance

Less than a month after Delphia and Global Predictions, the SEC announced an enforcement action sweep against firms for inappropriate performance advertising.

What constitutes "hypothetical performance" under the Marketing Rule is broad, however, some examples include backtested, projected, and model performance. Because hypothetical performance is theoretical, the SEC views it as potentially misleading. Consequently, the SEC expects advisers to accompany hypothetical performance with context, such as relevant assumptions made, criteria used, and related risks and limitations.

In the sweep, the SEC concentrated on firms that had posted hypothetical performance metrics to their general, public-facing websites. Public advertising of hypothetical performance contrasts with the Marketing Rule, which requires that firms advertise hypothetical performance only when it is relevant to the likely financial situation and investment objectives of the intended audience. By posting hypothetical performance to their public-facing websites, the advisers did not (really, could not) tailor it or limit its distribution appropriately. As such, these firms were censured by the SEC and ordered to pay civil monetary penalties.

The SEC also found violations in how one adviser depicted model performance on its factsheets. The adviser’s factsheets benchmarked model portfolio performance to the S&P 500. However, the factsheets showed the benchmark’s price returns, rather than total returns with dividends reinvested, which was how the model portfolio performance was calculated. This adviser also had other performance advertising deficiencies, such as presenting gross performance without also presenting net performance.

Takeaway

A straightforward takeaway from the hypothetical performance settled enforcement actions is that firms should generally not advertise hypothetical performance publicly. Because hypothetical performance was made available to mass audiences, the advisers could not form any expectations about their recipients' financial situation or investment objectives. The SEC observed in the Marketing Rule adopting release that, "We believe that advisers generally would not be able to include hypothetical performance in advertisements directed to a mass audience or intended for general circulation."⁵

The settled enforcement actions also demonstrate the SEC’s under-the-hood focus on performance advertising, including on benchmark construction and comparison. This is consistent with the SEC’s 2024 Risk Alert on the Marketing Rule, in which the SEC identified the following examples of misleading performance advertising:

• Benchmark index comparisons that did not define the index or provide sufficient information to enable an investor to understand the basis for the comparisons or disclose that the benchmark performance did not include reinvestment of dividends; and
   
• Performance presentations including outdated market data (e.g., market data that was more than five years old); and investment products that were no longer available to clients and reflected lower investment costs than the products that actually were available to clients.

Conclusion

These settled enforcement actions—from "AI washing" to hypothetical and model performance—evidence the SEC’s focus on Marketing Rule compliance. The increasing adoption of AI and other novel technologies will likely increase this focus. As always, advisers must ensure that their marketing materials accurately describe their business practices. New technologies are no exception. Advisers also must not neglect more traditional Marketing Rule compliance, including ensuring that all performance advertising abides by the Marketing Rule's strict requirements.^C

¹ Investment Advisers Act of 1940 Rule 206(4)-1
²In the Matter of Delphia (USA) Inc., Inv. Adv. Act. Rel. No. 6573 (Mar. 18, 2024), available at https://www.sec.gov/files/litigation/admin/2024/ia-6573.pdf
³In the Matter of Global Predictions, Inc., Inv. Adv. Act Rel. No. 6574 (Mar. 18, 2024), available at https://www.sec.gov/files/litigation/admin/2024/ia-6574.pdf
⁴SEC Charges Two Investment Advisers with Making False and Misleading Statements About Their Use of Artificial Intelligence, U.S. Sec. & Exch. Comm’n (Mar. 18, 2024), available at https://www.sec.gov/news/press-release/2024-36
⁵ Investment Adviser Marketing, Release No. IA-5653 (December 22, 2020) (effective May 4, 2021) at 201