The section of the investor update that takes longest to write isn’t the numbers. It’s the paragraph about AI. You know roughly what the business is doing. The question is how to say it in a way that sounds like progress without overstating what you can actually evidence.
That tension is real and worth taking seriously. The regulator has started treating AI overclaims as a securities matter, acquirers are forensic about AI claims during due diligence, and the personal exposure for anything misleading in an investor communication sits with you as founder, not with the technical team who built the system.
What counts as AI washing in an investor communication?
AI washing is making materially false or misleading statements about AI capabilities to investors. The US Securities and Exchange Commission has brought enforcement actions under this label, applying Rule 10b-5 of federal securities law to companies that exaggerated the role AI played in their operations. The standard is whether you had a reasonable basis for the claim when you made it, and whether you disclosed the limits alongside it.
The cases tend to share a shape. A company puts confident, forward-leaning language about AI in its investor materials, claiming the product is AI-powered, that machine learning sits at the core of the business, or that AI is central to operations. The underlying evidence does not match. Sometimes the system exists but plays a far less central role than described. Sometimes it barely exists at all. The Regulatory Review documents this in the SEC’s own words. Companies “have to be honest about the role AI plays in their business and not exaggerate it to the point of AI washing.” That language is now embedded in regulatory guidance, not just enforcement correspondence.
Why does the personal exposure land on the founder, not the technical team?
The founder signs off on investor communications. That is the starting point for where liability sits. Regulatory enforcement and corporate litigation have established that director-level accountability for AI-related claims does not transfer to the technical team that built the system or supplied the figures. Harvard Law Review research documents cases where boards have been held accountable for AI misrepresentations even when the specific claims originated with engineers or technical leads.
This matters particularly in founder-led businesses where the delegate running the AI programme often drafts the relevant section. A well-meaning delegate may write that the AI is having a major impact on efficiency because from inside the project, that genuinely reflects the feeling. But the investor communication goes out under your name, and if that claim cannot be evidenced to the regulator’s satisfaction, you carry the exposure. The technical team carries no obligation under Rule 10b-5.
The volume of SEC enforcement is worth registering. The SEC filed 456 enforcement actions during its 2025 fiscal year, many with AI-related components. That reflects a regulator that has moved from issuing guidance to actively pursuing cases. For founder-CEOs running investor-backed businesses, this is relevant regardless of whether the business is publicly listed, since the fraud provisions in securities law apply to any securities transaction.
Where in the investor update does this risk show up?
The risk shows up in the narrative section, not the financial tables. The paragraph or two where you describe what AI is doing for the business, that is where overclaims tend to live. Phrases like “we are now AI-driven”, “AI is central to our operations”, or “our AI is delivering significant efficiency improvements” invite scrutiny when the evidentiary base behind them is thin.
The SEC’s comment letter guidance makes the specificity bar explicit. The regulator has asked companies in approximately 61% of its AI-related comment letters to clarify “how the AI is or is intended to be used in those initiatives, projects, or technologies and any attendant risks.” That is a direct signal about what level of detail is expected when AI features in corporate disclosures, and it applies equally when you are writing to your own board.
Beyond regulatory exposure, there is the due diligence dimension. Morgan Lewis research from 2026 describes AI readiness as “a decisive factor in both competitive positioning and transaction value” in M&A processes. Acquirers now verify AI claims against internal documentation during due diligence. A gap between what was claimed in investor materials and what was actually built is a deal risk, and often a price-adjustment point.
What does having a ‘reasonable basis’ look like in practice?
Having a reasonable basis means you can point to documented evidence that directly supports each AI claim, and that you have disclosed the scope of that evidence in the same breath. If a pilot is showing a 30% reduction in processing time for one team on one task, that is a legitimate evidence base. The claim in the investor update should reflect that scope, not extrapolate from it across the whole business.
The practical difference looks like specificity rather than sweep. “Our AI pilot in invoice processing has reduced processing time by 30% across the finance team” is a defensible statement with a reasonable basis behind it. “AI is delivering significant efficiency gains across the business” is the same result with the scope stripped out, and it is the version that creates exposure. The SEC’s guidance is direct on this. Companies should “avoid overstating their AI capabilities and making misleading disclosures about AI use or functionality if they cannot be supported.”
Describing what you have built accurately, with its current limits named, is more convincing and more defensible than either understatement or sweep. A specific claim with disclosed limits reads as command to a board or investor that has encountered enough vague AI enthusiasm to treat it as a signal of weak evidence rather than strong performance.
What connects to this, and why it matters beyond the next update
The discipline of evidenced AI claims sits at the junction of investor communications, board-level AI governance, and acquisition due diligence. Private equity and venture capital firms now incorporate AI governance assessments into standard due diligence processes. NACD survey data from 2025 shows 62% of directors now set aside dedicated board time for AI oversight discussions. The investor update is one surface of a broader expectation that will only grow.
Fifty-four percent of North American venture capital and private equity firms anticipate restrictions on their own AI usage due to governance concerns, according to Ocorian research. That reflects how seriously the investment community is taking this as an operational risk, not a soft concern. For a founder-led business with an exit in view, the AI governance story you are building now, evidenced claim by evidenced claim, is part of what an acquirer will assess when the time comes.
The investor update where you get the language right is also the start of the paper trail that supports a clean exit. Evidencing AI claims is part of running the business well. Getting it right now means there is something substantive to show when it matters most.
