Delphia’s Robots Weren’t Everything They Claimed

Until recently, Delphia (USA) Inc. was an SEC-registered investment adviser based in Toronto. Delphia managed about $7 million for about 29,000 individual retail accounts through robo-advisory services and also $180 million for five pooled investment vehicles. But not anymore because the SEC sued it on Monday and they have moved the investment vehicles to a new adviser.

It All Started with So Much Hope

When Delphia started its robo-adviser business in 2019[i] it developed algorithms to manage retail client portfolios based on different investment objectives and risk profiles. While the algorithms were based on a variety of information sources, Delphia intended to use artificial intelligence and machine learning to collect data from its clients (such as from social media, banking, credit card, online purchases, etc.) as inputs into its algorithms. Delphia never quite realized this goal, though. While it did collect some client data intermittently, it never actually plugged that data into its algorithms.

Public Statements about Delphia’s AI

Here’s what they said:

  • In its Form ADV brochures Delphia claimed its advice was “powered by the insights it makes when individuals . . . connect their social media, banking, and other accounts . . . or respond to Delphia’s questionnaires” which make its investment decisions “more robust and accurate[.]” Delphia also said this client data was used in “a predictive algorithmic model” for the selection of “stocks, ETFs and options[.]”

  • In a December 2019 press release, Delphia claimed that it was “the first investment adviser to convert personal data into a renewable source of investable capital . . . that will allow consumers to invest in the stock market using their personal data.” Also, “Delphia uses machine learning to analyze the collective data shared by its members to make intelligent investment decisions.”

  • For almost three years Delphia’s website claimed that it “turns your data into an unfair investing advantage” and “put[s] collective data to work to make our artificial intelligence smarter so it can predict which companies and trends are about to make it big and invest in them before everyone else.”

Narrator: [They had not actually done these things.]

The Exam Staff Pokes Around

In July 2021 the SEC’s exam staff started asking questions. Delphia didn’t have good answers and admitted that it hadn’t used any of its clients’ data or created an algorithm to use it. So Delphia updated the statements in its Form ADV to state that “[c]urrently, [Delphia] does not use [client data] to make investment decisions for its Clients.” Good! Delphia did not identify the new disclosures in its Form ADV brochure summary of material changes section. Bad! Delphia also told the staff in October 2021 that it would review its other public statements, and hired a compliance manager and retained two outside compliance consulting firms.

Delphia also took some steps to correct statements it had made about its use of client data. Specifically, it told some investors that it actually didn’t have enough client data to be a valid data source for its algorithms.

Delphia Doesn’t Learn Its Lesson

But this was not the end of Delphia’s story. It unfortunately continued to make misrepresentations regarding its client data through August 2023. To take some examples:

  • It emailed investors that their data was “helping [Delphia] train [its] algorithm for pursuing ever better returns” and that Delphia “will pool your data with everyone else’s to power our algorithm.”

  • A Delphia social media post said its “proprietary algorithm uses the data being invested by our members, so we can make stock selections across thousands of publicly traded companies up to seven financial quarters in the future.”

  • A press release claimed that “Delphia’s proprietary algorithms combine the data invested by its members with commercially available data, to make predictions across thousands of publicly traded companies up to two years into the future.”

These statements also weren’t true. And yet Delphia had assured the exam staff that it was cleaning up its act?

Lessons

Don’t make up stories about your AI capabilities? Obviously? Delphia agreed to a $225,000 civil penalty and to negligence-based violations, including the Marketing Rule, the Compliance Rule, and Advisers Act Section 206(2).  

 

SEC Charges Two Investment Advisers with Making False and Misleading Statements About Their Use of Artificial Intelligence, SEC Press Release (Mar. 18, 2024)

In re Delphia (USA) Inc., Admin. Proc. File No. 3-21894 (Mar. 18, 2024)


[i] This is a settled case and wasn’t litigated, so what follows are just allegations, not facts.

Previous
Previous

Global Predictions Isn’t Really an AI Case

Next
Next

World Acceptance Corp. shows FCPA cases are sometimes really easy