This provision could ban AI regulation for 10 years

Director of UC’s Applied AI Lab speaks on advanced tech safeguards

Congress' proposed “big beautiful bill” has a provision that would ban states from enforcing regulation on artificial intelligence models and systems for 10 years. 

Cincinnati Edition recently spoke with Jeffrey Shaffer, director of UC’s Applied AI Lab, for insight on what the provision's legal implications could be at the crossroads of innovation and safety. 

Shaffer told Cincinnati Edition that regulation on generative AI is critical. “There definitely needs to be more guardrails and more security,” he said. “The question is: who does that? Do we do it at a more global level, a federal level or a more local level?”

Shaffer pointed out the lack of clarity in the proposed AI provision. “The problem here is that, with the moratorium, the 10 years would eliminate states from doing anything. And then there’s no discussion at the moment of what the federal rules would be,” he said. Using the example of financial services, he explained that broad federal regulation can aid states in choosing to be more restrictive or not on any given regulatory guideline. And in turn, it can be easier for the federal government to streamline its regulations by using states’ legislation as a working model. 

Shaffer also emphasized the importance of regulation in the business of AI chatbots. “It’s easy to take them off their rails. With a few prompt injections, you can get it to do nefarious things,” he said. 

“I don’t think we can trust the current guardrails in the current state of AI, especially around deepfakes, deep videos, voice, all of that. With a couple minutes of anybody on YouTube, I can imitate their image or the sound of their voice. I can write fake and malicious content very quickly,” Shaffer said. He also added that the danger is not just external, but also puts companies’ proprietary and internal data at risk. 

On top of that, it’s hard to tell what the state of AI will be in a decade. Shaffer said even AI experts have conflicting opinions on topics such as whether AI tools will continue to require such large amounts of energy to operate in the future. “The 10 year time period here is so long in the field where we are advancing things week by week,” Shaffer said. 

Shaffer pushed for understanding and transparency when it comes to both utilizing AI tools and regulating them. “We need to shore up a lot of these definitions — even starting with the definition of what AI is — so it’s clear what we’re putting up guardrails around.”

Listen to the full interview on Cincinnati Edition

Featured image at top: AI-themed overlay on person using a laptop. Photo/iStock

Related Stories

1

This provision could ban AI regulation for 10 years

June 12, 2025

The proposed “big beautiful bill” has a provision that would ban states from enforcing regulation on artificial intelligence models and systems for 10 years. The legal implications of the bill’s moratorium on AI regulation are a source of confusion across the political spectrum. Cincinnati Edition on WVXU recently spoke with director of UC’s Applied AI Lab Jeffrey Shaffer for insight on what the provision could mean at the crossroads of innovation and safety.

3

How to spot deepfake images

August 15, 2024

As images created by artificial intelligence become more widespread and harder to differentiate from real photos and videos, it’s important for people to be aware of them, a University of Cincinnati professor told WVXU.