The Biden administration is urging for increased accountability in artificial intelligence (AI) systems, calling for mandatory audits and clarity on liability within the AI supply chain. This push is outlined in a National Telecommunications and Information Administration (NTIA) report, which stresses the need for accountability tools and consequences for AI-related risks. NTIA Administrator Alan Davidson emphasizes that accountability is essential to unlock the full potential of AI, empowering stakeholders to hold developers and deployers responsible for any adverse outcomes.
Aligned with President Biden’s executive order on AI, the NTIA report recommends measures such as independent audits of high-risk AI systems and pre-release certification for AI models in sectors like healthcare and finance. Additionally, the report proposes the development of “AI nutrition labels” to enhance standard information disclosures and facilitate transparency in AI deployment. Collaboration between the federal government and the private sector is essential to strengthen AI risk management and improve accountability mechanisms.
As courts and lawmakers grapple with defining liability for AI-related harms, the NTIA suggests convening legal experts and stakeholders to navigate existing regulations. The report underscores the importance of third-party evaluation methods to ensure accountability, emphasizing the role of exposure and liability protection. Furthermore, the NTIA recommends enhancing the federal government’s capacity to address cross-sectoral AI risks and establishing national registries for high-risk AI deployments and adverse incident reporting, reflecting a comprehensive approach to AI governance and accountability.
