Davidson Wants NTIA to Focus on Pragmatic, Law-Based AI Solutions
NTIA Administrator Alan Davidson announced Wed. the launch of the agency's public consultation process related to its forthcoming report to President Joe Biden on the risks, benefits and regulatory approaches to AI foundation models, as directed in a Biden AI executive order (see 2310300056). Speaking at an event hosted by the Center for Democracy and Technology, Davidson said the report will focus on pragmatic AI policies rooted in technical, economic and legal realities of the technology. The Biden order gave the Commerce Department 270 days to get public input and deliver the AI recommendations. Davidson said.
Davidson’s speech came as Republicans on Capitol Hill warned the Biden administration to avoid what they view as European overregulation of AI that will preempt its potential. Overly burdensome regulations contemplated in the EO could “hinder” U.S. competitiveness globally, said House Commerce Committee Chair Cathy McMorris Rodgers, R-Wash., during a hearing Wednesday.
AI technology raises a number of “new concerns” since the passage of the EU’s general data protection regulation, testified Saif Khan, the Commerce Department secretary's senior adviser-critical and emerging technologies. Those concerns stem from newer AI models ingesting massive amounts of personal data, Khan said: This means there’s a role for new legislation, as well as technical expertise from agencies like Commerce. The computing power associated with AI creates new incentives for “data-greedy” developers and companies, testified Micky Tripathi, national coordinator-health IT, Department of Health and Human Services.
NTIA is focused on issuing guidelines for identifying AI-created material, including how to watermark it, said Khan. Another core issue is kids’ privacy, he said: There’s even more potential for mass data collection given the amount of time children spend on social media, Khan said.
Tech platforms should be liable for the real-world harms they do to children, including those related to mental health and suicide, said Sen. Josh Hawley, R-Mo., during a floor speech Wednesday. He failed to gain unanimous consent for his No Section 230 Immunity for AI Act (see 2312120075). Sen. Ted Cruz, R-Texas, was the only member to object on the floor. While he agrees with Hawley’s sentiment, Cruz said the bill hasn’t received proper vetting from the Senate Commerce Committee. Cruz is the committee's ranking member.
The bill removes government insulation for the platforms and allows parents to better protect their children, said Hawley. Cruz agreed Big Tech is doing “a lot of harm to our kids” and has been “complicit” in the greatest censorship of free speech in U.S. history. But unleashing trial lawyers to “sue the living daylight” out of Big Tech for every offense associated with social media harm isn’t the right approach, he said.