NTIA Mulls Ways to Audit AI Systems to Hold Models Accountable

The National Telecommunications and Information Administration (NTIA) is looking to create an auditing process for artificial intelligence(AI) systems, as a part of its efforts to promote safe and ethical use of the technology.

“I think one of the things that we've seen is, like financial audits for the financial accounting system, there is going to be a role to play for audits in the AI ecosystem,” Alan Davidson, Assistant Secretary of Commerce for Communications and Information, and administrator of the National Telecommunications and Information Administration, U.S. Department of Commerce said while speaking at a conference.

Davidson expressed that AI, particularly since the introduction of ChatGPT, has garnered significant attention, and ensuring the responsible use of AI can bring a ‘huge positive difference’ in people’s lives. However, to tap into the technology’s potential, Davidson believes tackling the risk associated with it is paramount.

He mentioned that his department is investigating the entire accountability lifecycle, beginning with the transparency and information necessary to hold models accountable. This encompasses considerations for the broader AI ecosystem, including determining who will conduct these assessments and exploring ways to establish a market of third-party auditors similar to what was done in the financial auditing system.

The NTIA, part of the US Department of Commerce, acts as the President's chief advisor on telecommunications policies along with overseeing the regulation of the telecommunications industry.

In April 2023, the NTIA sought public feedback on ways to mitigate AI risk and received roughly 1,447 written comments.

With the “Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence,” the Joe Biden-led administration mandated all developers of the technology as well as federal agencies to ensure AI developments are safe, secure, and trustworthy.