How A Founder’s AI Pitch Deck Can Become A Crime Scene

This article was originally published on Law360 in June 2026.

On April 17, federal prosecutors unsealed a 10-count indictment in the U.S. District Court for the Eastern District of New York charging the CEO and chief financial officer of iLearningEngines Inc., a corporate education and training company claiming to utilize artificial intelligence, withsecurities fraud, wire fraud, conspiracy and running a continuing financial crimes enterprise.[1] The criminal enterprise charge alone carries a maximum sentence of life inprison.

It was not an isolated case. It was the latest in a series of developments that began 13 months earlier andcontinues to unfold.

Many AI startup pitches follow the same choreography. The founder takes investors through the obligatory total addressable market slide, then arrives at the moment that closes rounds: the capability demonstration. The AI is faster, smarter, more autonomous than anything else on the market. The numbers are compelling. The investors lean forward. The checks are cut.

The pitch deck is now a crime scene. The lawyers reviewing it are already inside the tape.

The enforcement campaign against AI washing is no longer a prediction. It is a documented, accelerating fact. Private class actions alleging AI-related misrepresentation doubled between 2023 and 2024, and increased again in 2025.[2]

Apple Inc. lost nearly a quarter of its market capitalization from its December 2024 peak — erasing roughly$900 billion in market value at the trough — amid allegations that its public AI feature claims outpaced actual product readiness, and settled the resulting consumer class action for $250 million in May 2026; the stock has since recovered and now trades well above its prior peak.[3]

C3.ai faces a securities class action — Liggett v. C3.ai Inc., filed in the U.S. District Court for the Northern District of California in August 2025 — over alleged investor deception about AI performance.[4]

On Feb. 11, 2026, then-U.S. Securities and Exchange Commission Enforcement Director Margaret “Meg” Ryan publicly designated AI investor fraud a high-impact enforcement priority.[5] A couple of weeks later, the SEC updated its Enforcement Manual to treat AI washing Wells notices as routine workflow.[6] The SEC’s Cyber andEmerging Technologies Unit lists AI washing as an immediate enforcement target.[7]

And the U.S. Department of Justice has already brought its first criminal prosecution over AI-related misconduct. That prosecution is where this story begins.

Albert Saniger raised $42 million by following the familiar pitch choreography mentioned above.

On April 9, 2025, the DOJ indicted him for it in the U.S. District Court for the Southern District of New York.[8]Prosecutors there charged Saniger, the founder and former CEO of AI startup Nate Inc., with securities and wire fraud, alleging that the autonomous AI shopping technology he demonstrated to investors had an actual automation rate of effectively 0%. Thetransactions investors watched the app complete were being processed in real time by contract workers in thePhilippines and Romania. The pitch deck was a prop. The demo was a performance.

Thirteen months later, the Saniger indictment looks less like an isolated prosecution and more like the shot that opened a sustained campaign. Everything that has followed — the increased class action filings, Ryan’s Februaryremarks, the Enforcement Manual update, the SEC’s Investor Advisory Committee’s December 2025 call for standardized AI disclosure guidance — is the expansion of the infrastructure established by that case.

The legal, factual and political conditions for criminal prosecution of AI washing are not being assembled. They are already in place and actively in use.

The lawyers who review pitch decks are standing in the middle of all of it.

A New Kind of Exposure

For years, the framework for thinking about overstated AI claims in fundraising materials was the SEC’s civil enforcement model. A company oversells its technology, the commission opens an investigation, and the matterresolves in a settlement with no admission of liability and a manageable financial penalty.

The Presto Automation Inc. settlement in January 2025 followed exactly that script.[9] Practitioners noted it, tightened some disclosure language and moved on.

The Saniger indictment operates on a different legal register entirely. The Southern District of New York did notrefer the matter to the SEC and stand down. It brought parallel criminal charges; securities fraud under Title 18 of the U.S. Code, Section 1348; and wire fraud under Section 1343, running simultaneously with the civil action.[10]

Then-acting U.S. attorney of the Southern District of New York, Matthew Podolsky, made clear this was not anaccident of prosecutorial zeal. Saniger’s “deception,” he stated, “not only victimizes innocent investors, it diverts capital from legitimate startups, makes investors skeptical of real breakthroughs, and ultimately impedes the progress of AI development.”[11]

That framing has held across administrations and across agencies. Notably, the Saniger prosecution was the first AI washing criminal action brought under the Trump administration, a signal that enforcement posture on this issue is bipartisan and durable.[12] The DOJ and SEC are now described by securities defense practitioners as operating a coordinated enforcement posture in which exaggerated or unsubstantiated AI claims are treated as modern iterations of well-established fraud principles, not as industry buzzwords exempt from scrutiny.

The Saniger prosecution explicitly targets private market fundraising. Venture rounds, seed rounds and Series Acloses are within the Southern District of New York’s declared scope of concern. This is not a public company problem wearing a startup’s clothing. It is a startup problem, period.

A second enforcement signal arrived in December 2025, when the IAC formally recommended that the SECissue standardized AI disclosure guidance.[13] The IAC cited a “lack of consistency” in how companies describe AI capabilities to investors and called for

issuers to define what they mean by AI, disclose board-level oversight of AI deployment, and report separately on material AI effects on internal operations and consumer-facing products.

SEC Chairman Paul Atkins has been cool to new rulemaking, but the IAC’s framework is not merely aspirational. It describes the disclosure standard that enforcement is already applying. Prosecutors tested exactly thosecategories in the Saniger indictment: definitional precision, governance accountability and deployment specificity.

Where the Line Falls

Securities law has always made room for promotional enthusiasm. Vague, forward-looking statements that no reasonable investor would treat as a factual guarantee fall within the traditional puffery doctrine, and courts have been willing to treat claims like “our AI will transform the industry” as aspirational rather than actionable.[14]

That tolerance has not disappeared. But the Saniger case illustrates exactly where it ends.

The conduct alleged was not loose optimism in a pitch meeting. It was specific: It claimed automation ratesbetween 93% and 97%, sustained across multiple fundraising rounds over three years, backed by demonstrations staged to conceal manual processing.[15] Saniger allegedly directed employees to prioritize manually completing transactions initiated by investors during those demos to make the AI appear to work forthe precise audience whose money he was raising.[16] The company’s internal AI development, according to prosecutors, never achieved consistent functionality at all.

The operational line follows from that pattern. A claim crosses from permissible promotion into actionable misrepresentation when it is specific enough that a reasonable investor would rely on it, material to theinvestment decision, and false or misleading at the time it is made.[17]

What separates civil liability from criminal prosecution is scienter, meaning knowledge of falsity or reckless disregard for it. But recklessness is a lower bar than founders typically assume. A founder who repeats a capability claim that the engineering team has internally questioned is not making an honest mistake. That gapbetween the deck and the codebase is where prosecutors find their case.

The Lawyer in the Room

Outside general counsel who review fundraising materials occupy a position of growing exposure, and most have not yet recalibrated to match it.

American Bar Association Model Rule 4.1 prohibits a lawyer from making a false statement of material fact to a third person in the course of representing a client.[18] When counsel reviews a pitch deck and provides comments, or simply approves it for distribution, that document is entering the stream of investorcommunications. Counsel who identify red flags in capability claims and say nothing, or who approve materials they have reason to doubt, risk both professional discipline and potential aiding-and-abetting liability under securities law.

The standard is not one of technical expertise. Lawyers are not expected to benchmark AI model performance. They are expected to ask questions and to ask them in writing.

When a deck asserts a specific automation or accuracy rate, counsel should ask who produced that number, on what dataset and under what conditions. When a data room represents that the AI has been validated or isproduction-ready, counsel should ask what “validated” means in the company’s own documentation and whether any internal voices have challenged it. When the capability claims in investor materials seem to outrun the company’s development stage or technical headcount, counsel should probe that gap explicitly.

What counsel cannot do, in the enforcement environment that the Saniger prosecution and its successors have created, is treat the AI section of a pitch deck as boilerplate. That section is no longer a marketing flourish. It is a representation of material fact to prospective investors, and the standard of care for its review has risen accordingly.

A Checklist for the New Standard

The pre-pitch review that the current environment demands is not complicated. It is the set of questions a federal prosecutor would ask if the round went wrong.

Best practices include identifying every AI capability claim in the pitch deck, data room and written investor materials, including representations about training data, automation rates, model validation and deployment status. For each claim, document the evidentiary basis: who produced the number, on what data and whetherany internal technical personnel have disputed or qualified it.

Distinguish rigorously between current capabilities and anticipated capabilities. Forward- looking statements require specific cautionary language tied to known risks, not generic disclaimers recycled from the last deal.

Confirm that the engineers who own the AI system have reviewed and affirmatively approved the capabilityrepresentations in the fundraising materials. A discrepancy between what the technical team believes and what the deck says is the fact pattern that generates a criminal referral.

Document the review itself. Outside counsel should memorialize the questions asked, the representationsreceived and any claims that could not be corroborated. That record is not a shield against liability for a fraudulent underlying claim, but it is essential evidence of professional conduct, and, in the current environment, its absence is its own risk.

Closing the Loop

The choreography has not changed. Founders still stand before investors and click past the total addressable market slide to the capability demonstration. The pitch still builds to the same moment. But the legal context surrounding that moment has shifted in a way that most of the people in the room have not yet absorbed.

The iLearningEngines indictment and the Saniger prosecution before it established that AI washing is a criminalenforcement priority, not a regulatory formality. The IAC’s December recommendations confirmed that disclosure expectations are rising regardless of whether formal SEC rulemaking follows. Ryan’s Feb. 11 remarks made plain that the current SEC leadership views AI-related investor fraud as a high-impact priority warranting aggressive pursuit.

Private plaintiffs, watching the regulatory enforcement record accumulate, have followed with class actions against public companies from Apple to C3.ai. The infrastructure for prosecution — legal,factual and political — is not being assembled. It is already in place and actively in use.

Founders who overstate what their AI does are not just risking an SEC comment letter. They are risking an indictment. And the outside counsel who review those pitch decks without asking hard questions are no longersimply failing their clients. They are standing in a potential crime scene, and they do not yet know it.

That is the moment to recalibrate. Before the next deck goes out. Before the next round closes. Before the next demonstration is staged for a room full of investors who are writing checks based on what the AI is supposed to do.

Endnotes

Press Release, U.S. Dep’t of Justice, Eastern District of N.Y., “Former Chief Executive Officer and Chief Financial Officer of Nasdaq-Listed Company Charged With Operating a Continuing Financial Crimes Enterprise in Multi-Year Scheme to Defraud Investors and Lenders” (Apr. 17, 2026),https://www.justice.gov/usao-edny/pr/former-chief-executive- officer-and-chief-financial-officer-nasdaq-listed-company. The indictment charges Puthugramam “Harish” Chidambaran and Sayyed Farhan Ali Naqvi under 18 U.S.C. § 1962(c) (continuing financial crimes enterprise), 18 U.S.C. §§ 1348 and 1349 (securities fraud and conspiracy), and 18 U.S.C. §§ 1343 and 1349 (wire fraud and conspiracy).
See Cornerstone Research, Securities Class Action Filings — Full-Year 2025 Review (Jan. 28, 2026) (reporting 16 AI-related class action filings in 2025, a record high, up from 14 in 2024 and 7 in 2023),https://www.cornerstone.com/insights/press-releases/overall-size-of- securities-class-action-filings-reached-new-heights-in-2025/; see also NERA Economic Consulting, Recent Trends in Securities Class Action Litigation: 2025 Full-Year Review (Jan. 21, 2026) (reporting 17 AI-related filings in 2025).
Tucker v. Apple Inc., No. 25-cv-05197 (N.D. Cal. filed June 20, 2025). Apple’s stock declinedapproximately 25% from its December 2024 peak (a market cap of approximately $3.9 trillion), erasing roughly $900 billion in market value at the trough, as investors recalibrated expectations following announcements of delays to its Apple Intelligence features. Apple’s market capitalization has sincerecovered to approximately $4.27 trillion as of June 12, 2026, well above its prior peak.
Liggett v. C3.ai Inc., No. 3:25-cv-07129 (N.D. Cal. filed Aug. 22, 2025) (alleging materially misleadingstatements during February to July 2025). Earlier securities litigation against C3.ai arose from alleged misstatements about its partnership with Baker Hughes. See Reckstin Family Trust v. C3.ai Inc., No. 22-cv-01413 (N.D. Cal. filed Mar. 4, 2022).
Margaret Ryan, Director, SEC Division of Enforcement, Remarks to the Los Angeles County Bar Association, 56th Annual Securities Regulation Seminar (Feb. 11, 2026), https://www.sec.gov/newsroom/speeches-statements/margaret-ryan-02-11-26-remarks-los-angeles-county-bar-association.
Press Release, SEC, “SEC’s Division of Enforcement Announces Updates to Enforcement Manual” (Feb. 24, 2026), https://www.sec.gov/newsroom/press-releases/2026-20-secs- division-enforcement-announces-updates-enforcement-manual. The updated Manual, the first revision since 2017, includes enhanced Wells process protections and commits to annual future reviews.
Press Release, SEC, “SEC Announces Enforcement Task Force Focused on Climate and ESG Issues” (Feb. 20, 2025) (announcing the Cyber and Emerging Technologies Unit to replace the Crypto Assets and Cyber Unit); see also SEC Enforcement, https://www.dlapiper.com/en-us/insights/publications/ai-outlook/2025/sec- emphasizes-focus-on-ai-washing (senior CETU officials confirming AI washing as “an immediate priority” at the Securities Enforcement Forum West 2025, May 15, 2025).
United States v. Saniger, No. 25-cr-00157 (S.D.N.Y. indicted Apr. 9, 2025); Press Release, U.S. Attorney’s Office, S.D.N.Y. (Apr. 9, 2025), https://www.justice.gov/usao- sdny/pr/founder-and-former-ceo-ai-startup-indicted-securities-fraud-and-wire-fraud. Charges were brought under 18 U.S.C. § 1348 (securities fraud) and18 U.S.C. § 1343 (wire fraud).
In the Matter of Presto Automation Inc., SEC Release No. 34-102208 (Jan. 14, 2025) (Order Instituting Cease-and-Desist Proceedings), https://www.sec.gov/litigation/admin/2025/34-102208.pdf. The SEC alleged that Presto misrepresented its Presto Voice AI drive-through ordering technology as proprietary and as eliminating humanorder-taking, while in fact relying on workers located primarily in the Philippines and India to process the majority of orders.
See United States v. Saniger, No. 25-cr-00157, Indictment Counts 1–2. The securities fraud count is brought under 18 U.S.C. § 1348 (Sarbanes-Oxley securities fraud), which prohibits knowingly executing or attempting to execute a scheme to defraud any person in connection with a security and carries a maximum25-year sentence; the wire fraud count is brought under 18 U.S.C. § 1343. The parallel SEC civil action is SEC v. Saniger, No. 1:25- cv-02937 (S.D.N.Y. filed Apr. 9, 2025), charging violations of Sections 17(a) of the Securities Act of 1933 and 10(b) of the Securities Exchange Act of 1934 and Rule 10b-5 thereunder.
Statement of Acting U.S. Attorney Matthew Podolsky, quoted in DOJ Press Release (Apr. 9, 2025), supra note 8.
See Debevoise & Plimpton, “The SEC and DOJ Signal Continued Focus on AI Washing Under Trump Administration” (Apr. 17, https://www.debevoise.com/insights/publications/2025/04/the-sec-and-doj-signal- continued-focus-on-ai-wash.
SEC Investor Advisory Committee, Disclosure Subcommittee Recommendation Regarding the Disclosure of Artificial Intelligence’s Impact on Operations, approved at the Dec. 4, 2025 IAC Meeting, https://www.sec.gov/files/approved-artificial-intelligence- disclosure-recommendation-120425.pdf. The recommendation calls for issuers to (1) adopt and disclose a definition of “artificial intelligence”; (2) discloseboard oversight mechanisms; and (3) report separately on material AI deployments affecting internal operations and consumer-facing matters.
The puffery doctrine holds that vague, optimistic statements that no reasonable investor would rely on asstatements of material fact are not actionable under the securities laws. See, e.g., In re Worlds of Wonder Sec. Litig., 35 F.3d 1407, 1418 (9th Cir. 1994); ECA, Local 134 IBEW Joint Pension Tr. of Chi. v. JP Morgan Chase Co., 553 F.3d 187, 206 (2d Cir. 2009).
United States v. Saniger, No. 25-cr-00157, Indictment ¶¶ 12-15 (S.D.N.Y. 2025). The indictment allegesSaniger represented to investors that Nate’s app had an automation rate ranging from “93% to 97%” across multiple fundraising rounds spanning spring 2019 through December 2022. The actual rate, prosecutors allege,was “effectively zero percent.”
United States v. Saniger, No. 25-cr-00157, Indictment ¶ 18 (alleging that Saniger directed Nate employees to prioritize completion of transactions initiated during investor demonstrations to prevent investors from discovering the company’s reliance on manual labor). See CBS News (Apr. 13, 2025),https://www.cbsnews.com/news/former-nate-ceo- human-workers-instead-of-ai-fraud/.
A misrepresentation is material under the securities laws if there is a substantial likelihood that areasonable investor would consider the information important in making an investment decision. TSC Indus. Inc. v. Northway Inc., 426 U.S. 438, 449 (1976). Scienter for criminal liability requires knowledge of falsity orreckless disregard for the truth. Ernst & Ernst v. Hochfelder, 425 U.S. 185, 193 n.12 (1976). Recklessness, in the securities fraud context, means conduct that is highly unreasonable and an extreme departure from the standards of ordinary care. See SEC v. Platforms Wireless Int’l Corp., 617 F.3d 1072, 1093 (9th Cir. 2010).
ABA Model Rules of Prof’l Conduct R. 4.1 (2023) (“In the course of representing a client a lawyer shall not knowingly: (a) make a false statement of material fact or law to a third person “). Potential civil liability for outside counsel who approve or facilitate materially false investor communications may also arise under Section 10(b) of the Securities Exchange Act of 1934,15 U.S.C. § 78j(b), and Rule 10b-5 thereunder, and under aiding- and-abetting theories codified in the Dodd-Frank Act, 15 U.S.C. § 78t(e).