社会ガイド

AI Washing and SEC Enforcement

AI washing means overstating or inventing how a company uses artificial intelligence, and the SEC treats it as a form of misleading statement to investors or clients.

4 分で読めます
最終更新日 2026年9月15日

読み始めるクイズに答えてください

このページでは4 分で読めます

概要
ディープダイブ
戦略的影響
The Future of AI Washing and SEC Enforcement
現実世界の実装
リスクとガードレール
実装ロードマップ
探検を続けましょう
よくある質問

概要

The first SEC cases came in March 2024, when investment advisers Delphia and Global Predictions settled charges and paid $400,000 in combined penalties. It matters because any firm describing its AI in marketing, filings or pitch decks can face the same antifraud and advertising rules that cover every other claim it makes.

ディープダイブ

The SEC has brought AI washing cases using existing law. No AI statute was needed. On March 18, 2024, the SEC announced settled charges against two investment advisers. Delphia (USA) Inc. had said it used machine learning on client data to inform its investment decisions, and the SEC found it had not done so. It paid $225,000. Global Predictions Inc. had called itself the "first regulated AI financial advisor" and could not substantiate claims about AI-driven forecasts. It paid $175,000 and also settled other Marketing Rule problems. Both cases relied on the Advisers Act antifraud provisions and the Marketing Rule, which bars advertisements containing material claims an adviser cannot substantiate. Later cases expanded the theory. In June 2024 the SEC and federal prosecutors charged Ilit Raz, founder of the recruiting startup Joonko, with misleading investors about the company's AI and its customers. In January 2025 the SEC settled with Presto Automation, its first AI washing case against a public company, over claims about drive-thru voice ordering technology. In April 2025 the SEC and the Justice Department charged Albert Saniger, founder of the shopping app Nate, alleging that the promised automation was largely done by human workers. Under new leadership in February 2025, the SEC created the Cyber and Emerging Technologies Unit, and its stated focus includes fraud involving emerging technologies such as AI. The FTC has also pursued deceptive AI claims aimed at consumers. A common misconception is that AI washing only covers fake AI. The cases also involve exaggerating how much a system does, hiding human involvement, and failing to disclose reliance on a third party's technology.

技術的な洞察

To describe AI truthfully, match every public claim to evidence someone can inspect. First, keep an inventory that lists each model, its function, its inputs, whether it is built in-house or licensed, and where people review or override it. Second, write claims at the level of what the system actually does. "A classifier flags transactions for analyst review" can be substantiated. "AI protects your money" cannot. Avoid firsts and superlatives unless you have documentation that supports them. Third, disclose human involvement and outside vendors when they matter to the claim, as the Presto order shows. Fourth, keep substantiation files with test results, dates and model versions. Under the Marketing Rule, an adviser needs a reasonable basis for believing it can substantiate material facts if the SEC asks. Finally, route website copy, pitch decks, social posts and investor materials through the same review. Many problem claims appear first in informal channels.

戦略的影響

リスクと安全性

AI による壊滅的な被害も日常的な被害も、誰がリスクを理解し、誰が行動できるかにかかっています。

より明確な判決

国民と専門家のリテラシーは、強力な安全政策が政治的に可能かどうかを左右します。

誇大広告を打ち破る

明確な説明は、誇大広告、研究室の PR、曖昧な倫理劇場に囚われることを減らします。

The Future of AI Washing and SEC Enforcement

AI washing enforcement has continued across a change in SEC leadership, which suggests it is treated as ordinary fraud and disclosure enforcement rather than a policy priority tied to one administration. Expect cases to keep relying on familiar tools: the Marketing Rule for advisers, antifraud provisions for issuers, and criminal charges where investors were deliberately deceived. As AI features become standard, the question will shift from whether a firm uses AI to whether its claims about accuracy, autonomy and results hold up. Firms with documented, specific and modest descriptions are best placed for that scrutiny.

現実世界の実装

An adviser's website says its models 'learn from your spending data to pick stocks,' but no such data feeds the portfolio process. That is the kind of gap between claim and practice that was at the center of the Delphia case.

A startup tells investors its app automates a task with AI while contract workers do most of the work by hand. That was the pattern in the charges against Nate's founder.

A public company announces that its voice product handles orders without human help, but most orders need offsite workers to step in. The SEC's 2025 order against Presto Automation dealt with claims like this.

A compliance team rewrites 'our AI manages your portfolio' as 'we use a statistical model to screen securities; an investment committee makes final decisions.' That version is accurate and can be substantiated.

リスクとガードレール

能力が複雑になる一方で、実存的なリスクを SF として扱います。
高度な自律性の下での調整による表面製品の安全性を混乱させる。
英語以外や専門家ではない聴衆には、低品質の情報源しか提供されません。

実装ロードマップ

製品の危害、誤使用、制御不能/調整不良のリスクを分離します。
どのような証拠がタイムラインと重大度についてのあなたの見方を変えるかを尋ねてください。
マーケティング上の主張よりも、一次情報源と具体的な評価を優先します。
意識だけでなく、キャリア、政策、資金、スキルなど、行動経路を 1 つ特定します。

探検を続けましょう

Free newsletter

Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.

One email each weekday. Unsubscribe in one click. We never sell or share your address.

Test yourself

Take the AI Washing and SEC Enforcement quiz

Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.

クイズを開始する

Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation

よくある質問

What is AI Washing and SEC Enforcement?

AI washing means overstating or inventing how a company uses artificial intelligence, and the SEC treats it as a form of misleading statement to investors or clients. The first SEC cases came in March 2024, when investment advisers Delphia and Global Predictions settled charges and paid $400,000 in combined penalties. It matters because any firm describing its AI in marketing, filings or pitch decks can face the same antifraud and advertising rules that cover every other claim it makes.

Which two investment advisers settled the SEC's first AI washing cases in March 2024?

The March 18, 2024 settlements involved Delphia (USA) Inc. and Global Predictions Inc.

What were the combined civil penalties in the Delphia and Global Predictions settlements?

Delphia paid $225,000 and Global Predictions paid $175,000, for $400,000 in total.

Which marketing claim was associated with Global Predictions?

Global Predictions called itself the first regulated AI financial advisor and could not substantiate its AI forecasting claims.

What made the January 2025 Presto Automation case notable?

Presto was the first public company the SEC charged over AI washing. The claims concerned its drive-thru voice ordering technology.

Which rule did the SEC rely on against the advisers, barring advertisements with material claims the firm cannot substantiate?

The Marketing Rule, along with the Advisers Act antifraud provisions, was the basis for the adviser cases.

学び続ける