By GARRY RAYNO, InDepthNH.org
CONCORD — An attempt to regulate AI in the state is a bridge too far for the businesses that provide telecom, internet and similar services.
Sen. Rebecca Perkins Kwoka, D-Portsmouth, is the prime sponsor of Senate Bill 657 which attempts to bring some state oversight to AI.
The bill establishes an oversight position in the Attorney General’s Office, creates a civil right of action for certain uses of artificial intelligence, and creates a study commission to oversee the use of artificial intelligence in the state.
Perkins Kwoka told the Senate Judiciary Committee her bill is a starting point and welcomes input and suggestions to arrive at some point of agreement.
“This is not a perfect solution, it is a starting position,” she said. “I hope we can find something that works right.”
She said many businesses are concerned that AI could be used for misrepresentation and appropriations that need to be addressed.
The intent is not to limit AI but to protect businesses from someone impersonating them, Perkins Kwoka said, and the study commission should explore AI and how it could be used to influence elections.
The pace of AI development is greatly increasing, she said, and they should not let another year pass without taking some action in the state’s best interest.
But Maura Weston, of Cornerstone Government Affairs representing New England Connectivity and Telecommunications Association said her organization has significant concerns about the reach of the bill, saying its broad focus will have unintended consequences and will lead to unwanted litigation.
She said they are not opposed to regulation, but they prefer a more surgical approach and asked the committee to recommend it be killed.
A better starting point, Weston suggested, was the study commission with a broad representation of stakeholders.
That would be more appropriate than to jump to criminalization, she said.
Association members include Cox, Comcast, Charter, Beezeline and NESN.
The companies use AI in many ways, she said. They use it for customer service, content, personalization, developing new products and services, to manage and run broadband networks, and to troubleshoot with customers.
The definitions would clearly capture everyday uses under the section for civil action, Weston noted.
She noted the section under deceptive practices that involves misrepresentations, content that appears to be authentic and is material to a transaction, decision, or assessment; and its use is not disclosed.
She noted a prohibited conduct is the transmission of political advocacy when the companies use AI to monitor what customers watch and use to increase the customers experience, which would fall under this definition.
They want to give their customers the best experience from their services, Weston said.
Under the bill, the plaintiff is not required to prove by a reasonable doubt the content was intended to mislead.
Fines under the right to action would be between $1,000 and $10,000 but could be as high as $25,000 per violation for repeat offenders.
Committee member Sen. Debra Altschiller, D-Stratham, said she understands any kind or regulation for an industry gets a little squirrely, they do not want parameters on them.
She agreed a surgical approach would be good, but at this time AI is everywhere.
“It is a little hard to be surgical when you are covered everywhere,” Altschiller said.
She understands that the definitions give the businesses heartburn, she said, asking if the issue is disclosure.
Weston said her organization is not looking for space and is not concerned about disclosure.
For example, she said, Comcast uses AI to monitor the performance of the in-home network and radio equipment so the customer has the ultimate experience with the wifi product.
The breadth of uses of AI are astronomical, but this is a cautionary tale of criminalization of AI for uses beneficial to the customers, Weston said.
Altschiller noted the stakeholders are customers of other agencies; do they feel significant fear about how AI is used across all kinds of platforms?
To not disclose they use AI to deliver a service or ad product of interest would be misleading, she said. That is raising very basic customer protection, Altschiller said.
Weston said many of their services would be captured with the definitions in the bill.
“What gives us heartburn is this very broad, very significant cause of action,” Weston said.
She also warned the committee members that an executive order was issued in Washington prohibiting states from regulating AI and the president said those who violate it may have federal funds withheld.
Weston noted the state has been allocated $200 million to help build out broadband networks and passing the bill may put that at risk.
Altschiller noted other similar executive orders from the White House have been overturned in courts because they put federal funds at risk.
Weston said she is not judging the executive order just warning that those dollars could be put at risk.
“We will not bow to blackmail.” Altschiller said.
The committee did not make an immediate recommendation on the bill.
Garry Rayno may be reached at garry.rayno@yahoo.com.