SEC AI Washing Enforcement: Every Action, Every Penalty, and What Investment Firms Must Do Now

The SEC has made AI washing one of its highest-profile enforcement priorities. Between February 2024 and April 2026, the Commission brought enforcement actions against multiple investment advisers and companies for making false or misleading statements about their use of artificial intelligence. The penalties have ranged from $175,000 to over $1.8 million per action.

Understanding exactly what triggered each enforcement action — and the specific regulatory provisions cited — is the most practical guide to what registered investment advisers and broker-dealers must document and control.

The Enforcement Actions: A Complete Record

Delphia (USA) Inc. — March 18, 2024

Penalty: $225,000 civil penalty plus censure and cease-and-desist order

What Delphia claimed: Delphia marketed itself to retail advisory clients as using AI and machine learning to analyze customer transaction data and social media activity to make investment decisions. Its marketing stated the firm used client data to "predict which companies and trends are about to make it big" through AI.

What the SEC found: Delphia was not using AI in the manner described. Its actual investment process did not use the client data or AI capabilities represented in its marketing materials. The statements were false and misleading.

Violations cited: Section 206(2) and Section 206(4) of the Investment Advisers Act (fraud); Rule 206(4)-1 (the Marketing Rule) for false and misleading statements in advertisements.

Key takeaway: The SEC's first explicit AI enforcement action established that AI capability claims in marketing materials — including website copy, pitch decks, and client presentations — must be substantiated. The fact that Delphia did not intend to defraud clients was irrelevant. Negligent misrepresentation under Section 206(2) does not require intent.

---

Global Predictions, Inc. — March 18, 2024

Penalty: $175,000 civil penalty plus censure and cease-and-desist order

What Global Predictions claimed: The firm marketed itself as the "first regulated AI financial advisor" and claimed to provide "expert AI-driven forecasts." Its marketing included claims about AI providing "accurate" predictions.

What the SEC found: The "first regulated AI financial advisor" claim was unsubstantiated. The AI forecast accuracy claims were misleading. The firm's disclosures did not adequately address the limitations of its AI models.

Violations cited: Section 206(2) and Section 206(4) of the Advisers Act; Rule 206(4)-1.

Key takeaway: Superlative AI claims ("first," "most advanced," "most accurate") require the same substantiation as any other advertising claim under the Marketing Rule. The absence of a competing firm with the same regulatory status does not make "first regulated AI advisor" accurate — the claim was found misleading in context.

---

Rockwell Capital Management — February 2, 2024

Penalty: $1,602,089 in disgorgement and prejudgment interest (Rockwell Capital); $223,229 civil penalty (Brian Sewell, founder)

What Rockwell claimed: Rockwell raised $1.2 million for a cryptocurrency trading fund, representing to investors that the fund employed AI and machine learning technology for trading.

What the SEC found: The AI and machine learning technology did not exist. The fund was fraudulent. Sewell made false representations to investors about AI capabilities that were never implemented.

Violations cited: Securities Act fraud; Exchange Act fraud. This was a criminal fraud case, not merely a Marketing Rule violation.

Key takeaway: The most severe AI washing cases involve complete fabrication of AI capabilities to raise investor capital. These cases receive criminal referrals and carry disgorgement obligations that dwarf civil penalties.

---

Presto Automation — January 14, 2025

Penalty: $100,000 civil penalty

What Presto claimed: Presto marketed its "Presto Voice" product as an AI-powered drive-through ordering system. Public statements represented the system as highly automated with AI handling customer orders.

What the SEC found: Presto made misleading public statements about Presto Voice's AI capabilities and the degree of automation in its operations. The reality of the product's AI capabilities did not match the public representations.

Violations cited: Securities Exchange Act Section 10(b) and Rule 10b-5 (antifraud provisions for public company disclosures).

Key takeaway: AI washing exposure extends beyond investment advisers to any public company making material statements about AI capabilities. For public companies, AI capability claims in earnings calls, press releases, and SEC filings are subject to the full securities antifraud framework.

---

Nathan Saniger / Nate Inc. — April 2025

Status: Complaint filed; criminal charges by DOJ filed simultaneously

What Nate claimed: Saniger raised more than $42 million from investors by claiming Nate's shopping app used AI to autonomously complete online purchases. He represented automation rates above 90% to investors.

What the SEC found: Nearly all orders were manually completed by human workers in the Philippines and elsewhere. The claimed AI automation was essentially nonexistent. Saniger fabricated success metrics in investor communications.

Violations cited: Securities Act and Exchange Act antifraud provisions; wire fraud (DOJ charges).

Key takeaway: The scale of fundraising in AI washing cases has attracted parallel DOJ prosecution. AI washing that constitutes investor fraud is not exclusively an SEC matter — it can result in criminal charges independent of civil penalties.

---

The Common Pattern Across All Actions

Every SEC AI washing enforcement action shares a common structure:

1. A firm makes specific, affirmative representations about its use of AI in marketing materials, public disclosures, or investor communications

2. The actual AI capabilities are either nonexistent, materially different from what was represented, or cannot be substantiated

3. The misrepresentation influences investor decisions, client relationships, or capital raising

4. Existing antifraud and marketing rule provisions apply — no AI-specific regulation is needed

The SEC's enforcement posture is settled: AI capability claims are treated identically to any other material representation about an investment adviser's business, investment process, or product capabilities.

What Registered Firms Must Do Right Now

Review your website, ADV Part 2, pitch decks, client letters, and social media for any reference to AI, machine learning, algorithmic processes, or "intelligent" capabilities. Every claim must have a contemporaneous substantiation file that documents the factual basis.

Build substantiation files. For each AI capability claim you intend to keep, document: what the AI actually does, what data it uses, what its limitations are, how it was tested, and who internally validated the capability description. This documentation is your defense in an examination or enforcement inquiry.

Review disclosures for accuracy. Global Predictions' "first regulated AI advisor" claim was misleading not because it was factually false in isolation but because it created a misleading impression in context. Review AI-related disclosures for misleading implications, not just outright falsehoods.

Update your Marketing Rule review workflow. Any AI-generated or AI-related marketing content should go through the same pre-approval workflow as all other advertising. That workflow must include a substantiation check specific to AI claims.

Train your CCO and marketing team. The Marketing Rule applies to all statements in all media. Staff producing AI-related content need to understand that the same substantiation requirements that apply to performance claims apply to AI capability claims.

Document the review. In every enforcement action, the absence of a documented review process was a factor in the examination finding. Documenting your substantiation process — even for claims you ultimately remove — demonstrates a compliance culture that mitigates penalty exposure.