By: Joyce Jia

As explored in The LLM Public Offering: Why One S-1 Filing Will Reshape AI’s Governance, the anticipated IPOs of Anthropic or OpenAI will mark a constitutional moment for AI governance, the first time a frontier AI company must submit its business model to the full discipline of SEC disclosure. The natural question that follows is whether the SEC is institutionally prepared for that moment. On balance, the answer is yes, though with important qualifications. The agency enters the AI IPO era with a well-tested disclosure architecture, a clear institutional commitment to extending it to AI, and a track record of adapting to new technology sectors over time. What remains is the more difficult task of translating that commitment into sector-specific guidance.

A Framework Built to Adapt

The SEC’s core disclosure architecture rests on durable, technology-neutral foundations. Materiality doctrine articulated in TSC Industries v. Northway compels disclosure of information a reasonable investor would consider important in making an investment decision. That principle is broad enough to reach the material risks posed by frontier AI companies: accumulated deficits and timelines to profitability, training data legal exposure, critical dependencies on cloud and semiconductor infrastructure, and arguably the most consequential, the systemic risk that a failure or material change in a foundational LLM could propagate downstream across the ecosystem of agent companies and enterprise applications built upon it.

The SEC’s Commitment: A Path Forward

The SEC has signaled its intention to extend its disclosure framework to AI through sustained institutional engagement. The agency’s Investor Advisory Committee (IAC) has kept AI disclosure on its agenda since the Biden Administration, most recently issuing a recommendation on December 4, 2025, for the Commission to develop “comprehensive guidance” establishing an AI-related disclosure framework for issuers (“IAC’s Recommendation”). That recommendation drew on a panel discussion held on March 6, 2025, convening asset managers, governance experts, and AI practitioners. 

The IAC’s Recommendation centers on three pillars: requiring issuers to define “artificial intelligence,” disclose board oversight of AI deployment, and report separately on AI’s material effects on internal operations and consumer-facing products. Notably, the IAC recommended integrating these requirements into existing Regulation S-K items rather than creating a standalone AI disclosure chapter, recognizing that the existing framework is flexible enough to accommodate AI-specific risks without a full rulemaking cycle. As the cybersecurity disclosure precedent discussed below illustrates, that choice underscores the SEC’s proven capacity to extend its framework incrementally to emerging technology risks.

The Cybersecurity Rule Shows the Way

The SEC’s 2023 cybersecurity disclosure rule, which created Form 8-K Item 1.05 for material cyber incidents and Regulation S-K Item 106 for annual governance and risk management disclosure, built on over a decade of iterative, non-binding guidance: the Division of Corporate Finance’s 2011 Disclosure Guidance on cybersecurity, followed by the SEC’s 2018 Interpretative Guidance on Public Company Cybersecurity Disclosures. That progression, from staff guidance to Commission statement to final rulemaking, produced a disclosure architecture that companies were able to absorb with minimal friction and that investors found decision-useful. 

The AI disclosure gap calls for the same iterative approach: beginning with staff guidance clarifying how existing Regulation S-K items 101, 103, 106, and 303 apply to AI-specific risks, then evolving toward more targeted requirements as the technology and its disclosure challenges become better understood. Initiated now, that sequence could yield formal guidance well ahead of the next generation of frontier AI IPOs. 

A Strong Framework with Key Uncertainties

Research by the AI Disclosures Project at the Social Science Research Council, analyzing more than 7,800 Form 8-K filings by public companies on AI between 2022 and 2025, found that roughly two-thirds of all AI-related disclosures are positive in nature, while risk disclosures, model failures, and safety guardrail changes are systematically underrepresented. Most of these filings appear under Item 8.01, a voluntary catch-all category, suggesting companies are genuinely uncertain both where to report AI events and when they cross the materiality threshold. That uncertainty is precisely what targeted SEC guidance is designed to resolve, and it underscores why the IAC’s Recommendation for issuer-facing clarity is a necessary next step.

The SEC enters the AI IPO era better prepared than the headlines suggest. Its governing principles are sound, its institutional commitment is evident, and the path forward is well-lit by the cybersecurity precedent. The remaining question is one of timing: whether the guidance arrives before the first S-1 does, or after.

# SEC Disclosure # IAC’s Recommendation #AI Governance