AI Claims in Investment Fraud: The Bots in the Pitch and the Money in the Accounts

False AI claims in investment fraud persuade investors to trust technology the business does not have or cannot use as advertised. The pitch may promise a model that learns from client data, forecasts that identify opportunities ahead of the market, or trading bots that produce profits automatically. If you worked on the code, reviewed the marketing or handled the accounts, you may know where the promised capability disappears.

Here we follow an investment business’s AI claim from the advertisement to the underlying system and the money it attracted. Our broader account of false product and business capability claims follows the same problem through demonstrations, tests and commercial promises in other businesses.

What Is AI Washing?

AI washing means selling a false or misleading account of how a business uses artificial intelligence. In investment services, that account can describe how the firm selects investments, generates forecasts or manages a client’s money. The Securities and Exchange Commission used the term when it announced settled charges against Delphia and Global Predictions over their claims about using AI. The commission also warned public companies about misleading claims concerning their AI adoption that could influence investment decisions.

The claim gives us something specific to examine. A statement that client data improves investment selection raises questions about the data collected and the model that used it. A promise of automated trading raises questions about the code and the trades. An advertised service can be compared with what the business actually provided to clients.

What Did the SEC Find About the Advisers’ AI Claims?

The Securities and Exchange Commission said Delphia and Global Predictions misled investors about how they used AI. Its March 18, 2024 announcement described claims in filings, websites and other public communications, with findings under the investment adviser Marketing Rule.

Delphia’s Client Data Was Supposed to Make the Model Smarter

The Securities and Exchange Commission alleged that Toronto-based Delphia made false and misleading statements about using AI and machine learning that incorporated client data into its investment process. The statements appeared in SEC filings, a press release and its website from 2019 to 2023. The advertised advantage was that collective data would help the system identify promising companies and trends before other investors did.

According to the commission’s account of its order, Delphia lacked the AI and machine learning capabilities it claimed. Delphia agreed to pay a $225,000 civil penalty to settle the charges. For someone working with the data or investment process, the useful comparison is concrete: which client information reached the system, what the system did with it and what clients were told it could do. We see the head start: the marketing beat the technology.

Global Predictions Claimed to Be the First Regulated AI Adviser

The Securities and Exchange Commission alleged that San Francisco-based Global Predictions made misleading AI claims on its website and social media in 2023. Its claims concerned both AI-driven forecasts and its supposed status as the first regulated AI financial adviser. Global Predictions agreed to pay a $175,000 civil penalty to settle the charges.

The agency also described a false claim that Global Predictions offered tax-loss harvesting services, along with an improper provision limiting liability in its advisory contract. A person who reviewed the forecast may know different facts from someone who handled the service offering or the client agreement.

The Securities and Exchange Commission tied the false advertising claims to its Marketing Rule. That puts the copy itself beside the product evidence: the website language, social post or filing identifies what investors were led to believe. Our explanation of SEC complaints and orders distinguishes allegations from findings in those documents.

Privvy’s Promised Trading Bots and the Missing Trades

The Securities and Exchange Commission’s complaint against Privvy Investments’ founder alleged a crypto investment scheme that raised approximately $12.3 million. The founder, operating through the Cypress, Texas business, promoted AI-based bots that would buy and sell crypto assets across trading platforms to exploit small price differences. Investors heard that automatic stop-loss coding would limit losses to 3%.

The complaint quotes the public website’s invitation to investors. The promise reduced the entire investment decision to naming the desired profit:

The complaint alleged that the purported bots lacked both AI and stop-loss functionality. According to the complaint, only approximately $380,000 went toward buying crypto assets, without using bots, and those transactions generated no profit. Approximately $5.5 million went to Ponzi-like payments to investors. Other investor money allegedly paid for the founder’s house, gambling and personal expenses.

The returns had a separate display system. The agency alleged that the founder uploaded false balances to a web app and emailed fabricated account statements showing profits. For someone handling account reporting, the important comparison would be between the balance displayed to an investor and the transactions that supposedly produced it. The complaint alleges that the app showed earnings the trading never made.

The same complaint shows AI serving another purpose. When investors could not withdraw their money, the founder allegedly used ChatGPT to generate a letter from an invented audit company. It falsely said accounts had been transferred and required identity verification before balances could move to a liquidation trust. In the commission’s telling, the trading technology was missing, but AI was available to write the excuse.

AI Trading Bots and Promises of Certain Profits

Promises of certain profits turn a claim about technology into a claim about what will happen to investors’ money. The Commodity Futures Trading Commission’s January 2024 advisory on AI investment scams describes promotions for trading bots, trade-signal algorithms and crypto-asset arbitrage systems that promise unusually high returns. It explains that AI cannot predict the future or sudden market changes.

The advisory’s Mirror Trading International example describes a commodity pool run by a South African operator and promoted as using a proprietary bot to trade foreign currencies. The Commodity Futures Trading Commission describes fake customer accounts and balances created with MetaTrader demo accounts. Very little money was traded, while money from new investors paid earlier investors. Someone who knows how the account display was populated can connect the apparent trading success to the money that actually moved.

That comparison also matters when an investment presentation uses performance figures or backtests. A simulated result, a live trade and an account balance answer different questions. We examine which one the investor saw and which one the business actually produced.

Which Records Show What the AI Actually Did?

The useful evidence connects a dated investor claim to the data, system and transactions behind it. The securities fraud evidence trail brings together public statements and internal records that may sit in different teams. For an AI claim, we work through that comparison with the person who knows the operation.

The website, filing, pitch or social post supplies the promise. Its date lets us compare that promise with the model or code version in use at the time. If marketing said client data improved investment selection, we examine what data entered the process and how it affected the selection. Development records and reviews can show what had been built and tested when investors heard the claim.

Trading and payment records answer the money question. In the Privvy allegations, the code, crypto purchases and account statements told different stories about the same investment. We connect the promised strategy to actual orders, transactions and uses of investor funds, including the source of money paid out as returns.

The approval history supplies another connection. Who wrote the claim, who reviewed the technical description and what limitations reached management before publication can matter as much as the final advertisement. A marketing revision or response to a technical objection may explain how the investor’s version came to differ from the system the team knew.

What Can an Insider Explain About an AI Claim?

An insider can connect the public promise to a decision or limitation that investors could not see. You may know that the advertised data never entered the model, that a demonstration used a different process, or that the account balance came from a manual upload. We work with you to distinguish what you saw directly from what the documents allow us to test.

Different roles bring different parts of that account. A product or engineering employee may explain the code and the version used. Investment operations may explain the trades and payments, while marketing or compliance may know how the claim was approved and what reviewers questioned. You may be the person who can tie those records together.

For securities violations, an eligible SEC whistleblower can receive 10% to 30% of qualifying sanctions collected. Separately, a whistleblower case under the False Claims Act can pay 15% to 30% of what the government recovers for fraud against government money. We assess the conduct and the applicable program with you, including bringing evidence of false investment claims to the SEC.

You Know What Was Behind the AI Pitch

If you know why the advertised AI differs from the system your business actually used, speak with our whistleblower experts about that gap. We build the case with you around the technology, the investor communications and the money, and bring in securities fraud whistleblower lawyers who can work through the technical evidence. We stay beside you from that first account through the case.