From Identity to Architecture
The framework’s most direct engagement with the current policy moment came through its intersection with Utah’s State-Endorsed Digital Identity program, established by S.B. 260 in 2025 and S.B. 275 in 2026. I presented the Fiduciary Commons Companion Framework at the April 2026 SEDI Summit, arguing that SEDI answers the identity question correctly but leaves the data governance question largely open.
A citizen who presents a SEDI-compliant credential to access a government service has exercised meaningful control over that moment of verification. What happens next remains governed by whatever data architecture the state already has in place. In most states, that architecture is the problem. SEDI was designed to respond to data architecture that is centralized, aggregated, retention-heavy, purpose-indifferent and lacking enforceable citizen rights.
SEDI’s passage without a single dissenting vote in either chamber of Utah’s legislature demonstrates something important: A citizen-principal, rights-first approach to digital governance is politically durable across the partisan spectrum. The Fiduciary Commons statutes are designed to extend that approach from the identity credential layer to the full scope of government data relationships: tax records, benefits determinations, law enforcement data, health information and the AI systems that increasingly make decisions about all of these.
READ MORE: Identity is the new perimeter for state government security.
The AI Governance Gap
Nowhere is the architectural argument more urgent than in AI governance. Three sessions I presented at the Internet Identity Workshop in April 2026 traced the constitutional case from first principles to practical implementation, and the AI accountability gap drew the sharpest discussion.
GAAFA addresses the AI governance layer that the other two statutes do not directly reach. The premise is straightforward: VIDA limits what data an AI system can access but does not govern what an AI system can do with data within a compliant architecture. A government agency can build purpose-sequestered databases and deploy zero-knowledge credential infrastructure, and still run a black-box model that denies thousands of benefits applications in a single processing run with no mechanism for challenge, review or explanation.
The statute establishes tiered requirements by risk. Tier 1 systems, those making rights-affecting automated decisions about benefits, licensing or enforcement, must use interpretable, auditable decision logic; black-box models require demonstrated necessity and enhanced oversight. Tier 2 covers internal administrative AI. Tier 3 covers citizen-facing information services and requires clear disclosure that the interaction involves an AI system, with mandatory escalation pathways to human officials.
GAAFA also classifies autonomous AI agents deployed by government as secondary fiduciaries under PDTA, with the deploying agency jointly and severally liable for their breaches. This addresses a gap that state AI governance frameworks have largely ignored. When a government AI system gives a citizen incorrect guidance on eligibility or rights to appeal, and the citizen acts in reliance on that guidance, no identifiable human official made the error, no record of the decision logic exists and no private right of action currently reaches the deploying agency. The harm is real; the accountability is absent. GAAFA closes that gap directly.
LEARN MORE: Brace for the transformational impact of AI in government.
Security as an Architectural Outcome
For state CIOs and CISOs grappling with shrinking confidence metrics and expanding threat landscapes, the Fiduciary Commons framework offers a reframing that may be more useful than the standard prescriptions.
Purpose-sequestered databases, technically enforced rather than merely policy-restricted, change the security calculus fundamentally. A breach of a tax administration system cannot expose benefits records if the two data sets are not accessible through the same system. Zero-knowledge verification means credential checks generate no storable personal data at the verification point, eliminating that point as a breach vector entirely. Distributed citizen-controlled credential wallets mean there is no central identity repository to compromise.
The practical consequence: Breach scope becomes bounded by purpose. A security failure in one purpose-sequestered system exposes only the data collected for that specific purpose. The current architecture, by contrast, means a single compromised system can expose the full cross-agency profile of every citizen whose records touch it. That is not a policy failure, it is an architectural one. The same design choices closing that vulnerability also reduce cost: Credentials verified cryptographically across all compliant services eliminate redundant agency-by-agency identity infrastructure, and organizations that do not store personal data cannot be breached for it.
State technology leaders should sit with one question: If confidence in government data security is falling even as investment rises, what does that tell us about the source of the problem? If the architecture is the source, better security policies are not the solution. The Fiduciary Commons framework names the structural cause and proposes a legislative path to address it directly. The constitutional premises that make SEDI right — that identity belongs to the person, that government’s role is endorsement rather than ownership, and that fiduciary obligation is the correct legal framework for the government-citizen data relationship — do not stop at the credential layer. They extend to every context in which government collects, retains, uses or discloses personal data about the people it serves. That is the full scope of the problem. The framework addresses it in full.
More information on the Fiduciary Commons framework, including the full text of VIDA, PDTA and GAAFA, is available at fiduciarycommons.com.
