NTIA Report Urges Auditing Guidance to Build AI Trust, Accountability

High-risk artificial intelligence platforms should undergo independent audits to establish trust in AI solutions, according to the AI Accountability Policy Report that the National Telecommunications and Information Administration released on Wednesday.

To come up with the evaluations, the report recommended the creation of AI auditing guidelines via a partnership between the federal government and stakeholders. Also included in the AI accountability report are policy recommendations on support and regulations, NTIA said.

To support the creation of independent AI system audits, the report urged the federal government to provide additional assistance to such bodies as the U.S. AI Safety Institute and to fund the establishment of a National AI Research Resource.

NTIA Administrator Alan Davidson stressed the need for accountability to gain AI’s benefits fully. The agency’s recommendations in the new report will enable regulators and users of AI solutions to pin accountability on their developers while opening opportunities to harness the technology’s potential, he added.

Earlier in December, Air Force Secretary Frank Kendall, an advocate of accountability in AI use, pointed out at the Reagan National Defense Forum that the challenge with AI use is not enforcing regulations but making people behind the technology accountable for what it does. Holding them accountable for the misuse of technology is a way to build trust and confidence in AI, he said.