This advisory addresses two significant developments affecting nonprofit organizations: the U.S. Department of the Treasury’s proposed revisions to Form 990 and the governance, compliance, and operational risks associated with the increased use of Artificial Intelligence.
Treasury Announces Form 990 Transparency Initiative
On April 23, 2026, the U.S. Department of the Treasury announced that the Internal Revenue Service (IRS) plans to revise Form 990 to increase transparency, strengthen tax administration, and provide clearer reporting on certain activities of Section 501(c)(3) organizations. These are expected to be the first significant revisions to Form 990 since 2008.
What Is Being Proposed?
While the IRS has not yet released the specific language of the proposed revisions, Treasury’s announcement identifies the following areas for enhanced reporting:
- Government grants and contracts: Organizations receiving public funding may be required to provide more detailed information about the sources and uses of those funds.
- Fiscal sponsorship arrangements: Organizations may be required to provide clearer information about sponsored projects, including who operates the projects, who controls the funds, and how those funds are used.
Treasury stated that the proposed changes are intended to help detect misconduct, improve revenue classification, and reduce the risk of fraud, abuse, and misuse of taxpayer funds.
Although Treasury’s announcement did not specifically identify broader governance disclosures as a separate area of revision, the initiative underscores the importance of strong board oversight, fiscal management, and accountability over public and charitable funds.
What Happens Next?
Treasury and the IRS expect to issue proposed regulations and provide a public comment period before any reporting changes are finalized. No effective date has been announced.
Organizations and their advisors will have an opportunity to review and comment on the proposed requirements before they take effect.
What This May Mean for Your Organization:
- Government funding will likely receive greater scrutiny. Incomplete or inconsistent reporting may raise questions about how public funds were classified and used. Organizations should reconcile government grant and contract revenue across Form 990, audited financial statements, the general ledger, the Schedule of Expenditures of Federal Awards, grant records, and USAspending.gov data, when applicable. They should also maintain complete documentation supporting expenditure, including invoices, approvals, payroll records, allocation methodologies, procurement documentation, and evidence that costs comply with applicable award requirements.
- Fiscal sponsorship arrangements may require greater transparency. Agreements should clearly identify who controls the sponsored project and its funds, approves expenditure, employs personnel, and owns project assets. Organizations should document decision-making authority, oversight responsibilities, administrative fees, reporting requirements, and termination provisions.
- Directors and officers should take notes. Increased transparency may result in greater scrutiny of how charitable and public funds are controlled and used. Directors and officers could face accountability—and, in serious cases, potential liability under existing law—if they participate in, approve, or fail to exercise appropriate oversight over the misuse of charitable assets. This is not solely an organizational compliance matter; it is also a board-governance risk.
- Form 990 should receive meaningful governing-body review. Organizations should establish and document a review process involving finance, program leadership, legal counsel when appropriate, and the board or audit committee. Particular attention should be given to significant narratives, governance disclosures, related-party transactions, government funding, and fiscal sponsorship arrangements. Form 990 should be treated as a public accountability document, not merely a tax filing.
Artificial Intelligence Governance
AI is increasingly being used for grant writing, donor communications, financial analysis, program eligibility, case management, and administrative activities. Its use creates significant confidentiality, accuracy, bias, compliance, and reputational risks.
What This May Mean for Your Organization:
- Sensitive information may be exposed through public AI platforms. Employees may enter donors, employees, clients, patients, students, beneficiaries, or financial information into AI tools without authorization. Organizations should adopt a written AI policy defining approved tools and permitted uses, prohibited information, activities requiring approval, human-review requirements, and incident-reporting procedures.
- AI-generated information may be inaccurate, incomplete, biased, or fabricated. These outputs could affect grant reports, financial statements, board materials, donor communications, employment decisions, or beneficiary eligibility. Organizations should require qualified human review and prohibit AI from serving as the final decision-maker in high-risk financial, compliance, employment, or programmatic matters.
- Unapproved “shadow AI” may bypass established controls. Organizations should assess how employees currently use AI, maintain an inventory of approved tools and use cases, provide practical training, and establish a process for requesting approval of new tools.
- Third-party AI vendors may create privacy and cybersecurity risks. Vendors may retain organizational information, use it to train their models, or rely on subcontractors without adequate control. Organizations should perform due diligence addressing data ownership, retention, model training, access controls, breach notification, data deletion, and contractual protections.
- Responsibility for AI oversight may be unclear. Organizations should assign an accountable executive, establish appropriate board or committee oversight, and periodically report on significant AI uses, incidents, and emerging risks. The AI policy should be reviewed at least annually and whenever significant new AI use is introduced.
