AICPA ET §1.400 and AI-generated marketing for audit firms

The AICPA Code of Professional Conduct does not say the word "ChatGPT." It does have a great deal to say about advertising that creates "false, misleading, or deceptive" expectations about audit services, and about independence obligations when marketing references specific clients.

Audit firms face a structurally harder version of the legal marketing problem. The independence rules are bright-line prohibitions on certain content patterns — not the kind of gray-area judgment Rule 7.1 affords. And the rules apply to anything attributable to the firm, including content a model rendered from your published material.

The two rules that actually bite

§1.400 — Advertising and other forms of solicitation

Section 1.400 of the AICPA Code prohibits advertising or solicitation that is "false, misleading, or deceptive." The interpretive guidance lists three categories that are particularly relevant to AI search output:

• Creates unjustified expectations of favorable results. "Helped clients save millions in tax" rendered as "guaranteed tax savings" is the canonical example.
• Implies the ability to influence regulators, courts, or officials.
• Compares services in a way that cannot be substantiated.

§1.600 — Independence

Marketing content that names specific audit clients — even with permission, even truthfully — crosses into §1.600 territory when it could be construed as creating a relationship that impairs independence. The classic case: case studies. The subtle case: when a prospective audit client is told by an AI engine that your firm "specializes in serving" a named company you actually audit.

The AI-specific risk: cited-client surfacing

LLMs are trained on aggregated public content. If your firm has a public case study about an audit engagement, the model has ingested it. When asked, the model may proactively name that client in unrelated contexts — surfacing the relationship in ways the original case study did not intend.

Cognoverge has seen audit firms whose model-rendered descriptions name 6–8 audit clients the firm never highlighted in marketing — the model assembled the picture from registration statements, proxy materials, and press releases. This is not the firm's fault, but it is the firm's reputational exposure and, in some jurisdictions, an independence concern when the named relationship is current.

The control framework

1. Standing language in every published case study

Every case study should include standing language about the scope of services and the timeframe of the engagement — phrased so it survives paraphrase. "Cognoverge & Co. served as the independent auditor of Acme Industries for fiscal years 2020 through 2023." Past-tense, scoped, dated. The model will re-render this with the past tense intact most of the time. Without standing language, the model defaults to present tense.

2. Independence disclaimers in firm overview content

Firm overview pages should explicitly note that "client lists identified in this overview reflect public information" and link to a current client-list policy. This both grounds the model and creates a defensible position if an independence concern is raised based on AI-surfaced relationships.

3. Cross-functional review of marketing content

Marketing claims about audit and assurance services should clear independence review, not just legal review. The independence concern is more likely to surface in AI-rendered content than in human-written copy because the model aggregates across data sources marketing never coordinated.

4. Submission to AICPA-aware index sources

Industry-specific directories (AICPA's own member directory, regional CPA society listings, audit committee resource portals like NACD) are weighted heavily in audit-related AI queries. Getting your firm's structured listing into these — with clean, scoped, paraphrase-safe descriptions — is high-leverage.

What the SEC overlay adds

For audit firms with PCAOB-registered engagements, Rule 2-01 adds a layer of independence prohibitions on relationships with audit clients. Marketing content that names a current audit client — including through AI paraphrase — should pass through independence review with the SEC overlay applied. The pattern we have seen at audit firms with strong AI-search practices: an independence-aware lint runs against marketing drafts the same way a legal lint runs against representation claims.

Why this matters more, not less, for AI marketing

Audit firms have historically used marketing as a secondary channel — the primary acquisition path is relationships, referrals, and RFP responses. AI search inverts that. When a board member or audit committee chair asks ChatGPT or Claude for a shortlist before issuing an RFP, the firm with strong model representation enters the RFP with an advantage that relationships alone cannot match.

Independence-aware AI marketing is not a defensive cost. It is the qualifier that lets an audit firm compete in AI search at all. Firms that ignore the rules will get punished — by the regulators if the content is non-compliant, by the model if the firm refuses to publish at all.