Federal Agencies Report Over 1,700 AI Use Cases Amid Transparency Challenges

Federal agencies have submitted their complete inventories of artificial intelligence (AI) use cases to the White House, marking a significant step toward ensuring the trustworthy and transparent deployment of this technology. Following the recent deadline on December 18, 2024, a total of 37 agencies reported over 1,700 use cases to the White House Office of Management and Budget, adhering to a directive issued by President Joe Biden in his October 2023 executive order on AI.

Overview of AI Use Cases

The directive required agencies to provide detailed information about how AI is influencing their operations and the measures taken to mitigate associated risks. Among the reported use cases, the top three categories identified were mission-enabling and operations support, health and medical system support, and government service operationalization. Out of a total of 1,757 reported use cases, 227 were noted as having implications for civil rights or safety.

Agency Highlights

The Department of Health and Human Services (HHS) led in reported use cases, with 271 active AI systems, four of which are identified as impacting civil rights. Following closely, the Department of Veterans Affairs reported 229 active use cases, including 145 classified as rights-impacting at various levels. The Department of Homeland Security and the Department of the Interior also reported significant numbers, with 183 and 180 active AI use cases, respectively.

Addressing Documentation Challenges

While the increased transparency surrounding AI usage is welcomed, analysts have pointed out persisting inconsistencies in how agencies document their AI practices. Quinn Anex-Ries, a senior policy analyst at the Center for Democracy & Technology, stated in an email, ‘Agencies’ updated AI use case inventories reflect both increased usage and improved documentation. Some inventories contain significantly more information than in previous years, marking important progress. However, there are still significant inconsistencies in documentation practices and limited data on risk evaluation.’ She emphasized the opportunity for agencies to learn from one another to improve future inventories.

Conclusion

In conclusion, as federal agencies seek to navigate the evolving landscape of AI deployment, the challenge remains to enhance transparency while consistently documenting AI usage across various sectors.