AI Governance Policy for US Agencies

On March 28, 2024, the White House moved to ensure the responsible development and use of artificial intelligence across the federal government. Vice President Kamala Harris announced the Office of Management and Budget issuance of the first-ever government-wide policy framework to mitigate the risks and harness the benefits of AI systems deployed by federal agencies.

This policy delivers on a key component of President Biden’s landmark Executive Order on AI issued last year. It sets the U.S. government on a path to become a global model and standard-bearer for safe, equitable and trustworthy AI innovation.

Under the new rules, federal agencies will be required to implement robust safeguards and risk management practices when using AI technologies that could impact American citizens’ rights, opportunities or safety. This includes mandatory actions like rigorous testing, monitoring for discriminatory impacts, providing transparency to the public, and ensuring human oversight on high-stakes decisions. As presented, the policy isn’t aimed at halting AI progress, but rather accelerating it responsibly. It directs agencies to embrace cutting-edge AI capabilities like generative AI while putting adequate guardrails in place. Agencies will also be encouraged to grow their AI workforce and technical expertise.

A major focus is bringing much-needed transparency to the government’s use of AI systems. Agencies will have to publish detailed annual inventories disclosing their AI use cases, potential public impacts, and mitigation strategies. Sensitive use cases will be reported internally with metrics. To drive accountability, agencies will be required to designate Chief AI Officers to coordinate AI governance and to create senior-level AI Governance Boards to provide oversight.

The impacts may be far-reaching. The policy aims to protect citizens’ civil rights, prevent algorithmic discrimination, strengthen AI procurement practices, boost public trust and open the door for AI-powered advances in areas, such as healthcare, climate change response and aviation safety.

The OMB policy provides concrete details to the Biden Administration’s vision laid out in its AI Bill of Rights for citizens. It builds on technical AI risk management resources developed by the National Institute of Standards and Technology.

With this policy, the White House is intending to stalk out a global leadership position on AI governance as other nations are grappling with crafting their own regulatory frameworks. If executed effectively, it could become a model that raises the bar for responsible AI practices economy-wide. Of course, the real test will be in the implementation across the federal bureaucracy. But this robust governance framework is a promising and principled start to harnessing AI’s awesome potential in the service of the public good.