AI washing is the practice of claiming AI-driven automation or intelligence for products that are actually powered primarily by human labor or conventional software. The government pursued criminal charges against Nate because the founders allegedly made knowing misrepresentations to investors about the company's AI capabilities while internally understanding those claims were false — meeting the elements of wire fraud, not just civil securities violations requiring only a material misstatement.
AI washing describes the practice of claiming artificial intelligence capabilities that a company does not actually have, or materially overstating the degree to which a product relies on AI rather than human labor. The practice had been widespread enough in the 2022–2024 period that the SEC issued guidance on disclosure obligations for AI-related claims and brought a handful of civil enforcement actions. The Nate case — United States v. Albert Saniger — was different in kind. Nate was a startup claiming to automate online checkout using AI. Federal prosecutors alleged that the company's "AI" was substantially performed by human contractors in the Philippines while the company raised capital from investors who believed they were funding an AI product. The criminal charge — wire fraud — carries significantly higher stakes than civil SEC enforcement: substantial prison exposure and a felony conviction rather than a fine and disgorgement. Gurpreet S. Bal's read on the prosecution is that the AI washing category had become large enough to require a deterrence case. "Once the government shows it's willing to go criminal on AI fraud, the entire population of startups that have been aggressive in their AI claims has to reassess their exposure," he says. "The civil enforcement actions were an annoyance for most companies. A criminal case gets the attention of boards and investors in a completely different way." The cases that followed — CaaStle, SKAEL, IRL, Done Global, AllHere, ComplYant — varied in the nature of the alleged misconduct, but they shared a common thread: venture-backed companies that had raised substantial capital on the basis of claimed metrics or capabilities that prosecutors alleged were fabricated or materially misrepresented.
Academic research on startup fraud consistently shows that it originates from concentrated founder control without independent oversight, pressure to hit investor-promised milestones with inadequate board scrutiny, and a culture that treats growth metrics as the primary measure of success regardless of how they are achieved. Companies where the founder controls the board and serves as both CEO and chairman, without strong independent audit mechanisms, are statistically most vulnerable to fraud escalation.
A 2025 empirical study of 614 venture-backed startup fraud cases — the largest systematic analysis of VC fraud to date — produced findings that challenge the intuitive assumption that fraud in startups is primarily a founder personality problem. The study found that founder-controlled boards — boards where the founding team retains voting control or board control — were associated with an 88% higher likelihood of fraud compared to companies with more balanced governance structures. More striking, governance variables predicted the incidence of fraud better than founder background characteristics: whether founders had prior criminal records, academic credentials, or serial entrepreneurship experience was less predictive than whether the company had functional independent oversight of the founding team. The implication is both reassuring and uncomfortable for the VC industry. Reassuring because it suggests that fraud is structurally preventable through governance design, not just a random outcome of bad actors. Uncomfortable because VC investors have historically advocated for — and in many cases required — founder-controlled governance structures as a condition of investment, on the theory that founder control produces better outcomes by allowing fast decisions and long-horizon thinking. "The data say that founder control, without independent oversight, creates conditions for fraud," Gurpreet S. Bal notes. "That is not an argument against founder control generally — it is an argument for what the counterweights need to look like." The separate but related VC litigation dataset, covering active venture firms between 2014 and 2025, found that approximately 25% of active VC firms had been involved in at least one lawsuit during the period, with fiduciary duty claims representing roughly 40% of the final period's cases. Litigation between investors and founders — once relatively rare in Silicon Valley, where reputational norms historically discouraged it — has become a routine feature of the asset class.
Co-founder disputes in 2026 increasingly involve allegations of IP misappropriation, corporate opportunity diversion, and evidence destruction — not just equity disputes. The Lux Optics case established that founders who delete messages or documents after litigation is reasonably anticipated face spoliation sanctions that can be outcome-determinative. Courts are applying adverse inference instructions aggressively, and the lesson is that digital communication hygiene matters as a legal risk before any dispute materializes.
Lux Optics v. de With was one of the most widely followed co-founder disputes of 2025–2026. The case involved allegations by the company against its departing co-founder of data destruction: 500 gigabytes of company data allegedly deleted in the period surrounding the co-founder's departure, along with disputed IP assignments and conduct involving a third party with Apple connections. The case illustrates a structural pattern that Gurpreet S. Bal has described as the new normal in co-founder disputes: what begins as an internal governance conflict over roles, equity, or strategic direction escalates into formal litigation involving IP ownership, fiduciary duty claims, and in some cases criminal referrals under statutes like the Computer Fraud and Abuse Act. "The data deletion pattern comes up in dispute after dispute," he says. "And it almost never helps the person who does it. Courts are not sympathetic to parties who spoliate evidence. It turns a complicated dispute into one where one side looks much worse." For investors, the co-founder litigation wave has a second dimension that the data capture: fiduciary duty claims against VC investors — alleging that investor board members acted in their own interest rather than the company's — are a growing share of VC-related litigation. This is consistent with the broader governance environment described in the companion piece on conflicts of interest in startup counsel: the economic relationships that tie VC funds to their portfolio companies create conflict structures that litigation is increasingly being used to test.
AI washing enforcement creates liability exposure for board members who approved or failed to correct misleading investor materials, and for lead investors who were aware of discrepancies between the company's public claims and its actual capabilities. The duty to not mislead investors extends to directors and control persons, not just the CEO who signed the communications. Board members should expect that investor communication review is now a governance obligation, not an optional courtesy.
The prosecution of VC-backed founders for AI washing claims creates an underappreciated exposure for the board members and investors who received and approved the communications at issue. Wire fraud conspiracy liability under federal law does not require the government to prove that a co-conspirator personally made the false statement — it requires proof that they knowingly participated in the scheme. A board member who received a pitch deck claiming AI capabilities and approved a fundraising round built on those claims without reasonable inquiry has a different exposure profile than one who raised questions, documented concerns, or required an independent technical audit. "The investor and board dimension of these cases is still developing," Gurpreet S. Bal notes. "The prosecutions so far have targeted founders. But the inquiry process, the documentation of diligence, the level of scrutiny applied to AI capability claims — all of that will matter if prosecutors look at whether there were other knowing participants." For investors currently on boards of AI startups, the implication is concrete: ensure that the AI claims in investor materials have been reviewed by technical staff or outside advisors, document that review, and ensure that any representation to new investors accurately characterizes what the product actually does. The SEC has also signaled continued focus on AI-related disclosure obligations in the public markets, and pre-IPO companies going through S-1 preparation should expect detailed scrutiny of any AI claims in their registration statements.
Founders should implement a disclosure controls policy requiring legal review of any AI capability claims before publication, maintain board minutes that reflect genuine oversight of company-stated metrics, and establish a clear policy for data preservation that prevents document destruction when any dispute or regulatory inquiry is anticipated. Investors should conduct reference checks on AI capability claims during diligence and build representation and warranty provisions that specifically address AI performance claims in their term sheets.
The pattern across the 2025–2026 cases is consistent enough to support a clear set of practical responses. First, AI capability claims in investor materials, customer contracts, and press releases should reflect what the product actually does today — not what it is designed to do, not what a future version will do, and not what human-assisted performance looks like if marketed as automated. The gap between "AI-assisted" and "AI-powered" may seem like marketing language; it is increasingly the line prosecutors and SEC enforcement staff are drawing. Second, the governance data argues for board structures with meaningful independent oversight, even at companies where founders prefer control. Independent board members who ask hard questions about product capabilities, revenue recognition, and financial performance are not a governance burden — they are a fraud prevention mechanism, and the academic data confirm this with a level of statistical significance that is difficult to dismiss. Third, for investors, the fiduciary duty litigation trend is a reason to be deliberate about how board decisions are made and documented, particularly in situations — down rounds, acquihires, founder exits — where the investor's interests and the company's interests might diverge. A board process that is careful, documented, and properly attentive to all shareholders' interests is far harder to challenge in litigation than one that appears to have been managed for a single constituency. See the companion piece on founder pushouts and co-founder disputes for the specific governance and legal dynamics of the situations most likely to generate litigation exposure.
Gurpreet S. Bal is a corporate partner with 16 years advising on private equity, merger transactions, and public offerings for companies and investors at three of the world’s top law firms. He has represented clients in hundreds of transactions with aggregate deal value exceeding $60 billion across AI, semiconductors, fintech, and emerging technology. For more information and to get in touch, visit gurpreetbal.com.