Pennsylvania will continue to prosecute artificial intelligence laws for the benefit of its residents in response to the Trump administration's executive order violating state-level AI regulations, state Attorney General Dave said last Sunday.
“The reality is that the executive order is, for lack of a better term, somewhat vague, and we're going to have to sit back and watch for a little while and see what that actually does in Pennsylvania,” he said Sunday.
It will be difficult to know what the state's response will be until the newly created federal task force takes action, but “we are clearly opposed to pre-empting state regulations,” Sunday said.
“If there is a law that is attacked on grounds of constitutionality, we will absolutely defend it,” he said.
As attorney general, Sunday will be tasked with defending Pennsylvania law in court, including a series of new and pending AI-related laws. Many proposals focus on notifications, such as a proposal to add AI-generated images to the definition of child sexual abuse materials, alongside a pending bill to be considered by the Technology Committee on Monday.
Sunday said he would “absolutely” recommend that prosecutors continue to pursue these criminal charges related to the use of AI “deepfakes” to create pornography, for example, regardless of the executive order.
“We think it's important that we have the ability to prosecute people who use this tool to commit horrific, horrific acts that harm people,” he said Sunday. “We must be in a position to hold these bad actors accountable.”
presidential order
President Donald Trump signed an executive order earlier this month overriding state laws, citing the need to create national policy. A new AI Litigation Task Force could file legal challenges to states with AI regulations, with exceptions for child safety and data center infrastructure.
In particular, President Trump said AI companies need one set of rules for approval or disapproval, rather than “50 different sources.”
“I understand the thinking behind it,” Sunday said.
“But at the same time, we all have a job to do. I'm the attorney general of Pennsylvania, and my job is to do everything I can to protect the public, keep people safe, and keep our children safe. So I'll always advocate that we should do that in my job,” he said.
Earlier this month, Sunday led a bipartisan effort by 42 attorneys general to pressure major AI companies to identify guardrails to reduce AI sycophancy, “when artificial intelligence models single-mindedly seek human approval,” and delusion.
The Dec. 9 letter cited real-world deaths (including murders and suicides), mental illness hospitalizations, and “other delusional spirals” related to the use of AI, and asked companies to respond by Jan. 16.
“AI can do amazing things, and we have to embrace that,” he said on Sunday. “That's why we, as attorneys general, have decided to go directly to the companies and … hopefully work with them to put them in a position to put these guardrails in place.”
Sunday said the attached letter was not written “in response” to the anticipated executive order.
Capitol AI
Meanwhile, state lawmakers are considering legislation that would require health care providers and insurance companies to notify customers about their use of AI, explicitly citing a lack of federal regulation as a motivating factor.
“There is no real federal regulation at this point,” the bill's sponsor, Rep. Arvind Venkat (D-Allegheny), told the House Communications and Technology Committee last week.
In particular, Venkat was concerned about the “black box” of AI, which makes clinical decisions that cannot be consistently reproduced because the algorithms used are closely guarded trade secrets.
Venkat's House Bill 1925 received support from multiple stakeholders seeking human oversight, including Sunday Office representatives, consumer advocacy groups, and providers.
Emergency physicians like Venkat spoke about the potential benefits of AI. WellSpan Health's David Vega said the technology presents an opportunity to “bring the humanity back to health care,” citing the tendency of electronic health records to turn health care providers into “data clerks” taking notes.
“I see the implementation of AI as a success. It holds the promise of shifting our focus back to the patient. We can stop looking at a computer screen and instead spend our time talking to patients,” Vega said.
Other providers noted that the bill extends reporting requirements to third-party contractors, even though providers may not know about their use of AI.
Jonathan Greer, president and CEO of the Pennsylvania Insurance Federation, pointed to the prioritization of electronic records as a reason for the federal government's intervention in AI. Before the national law was enacted, each state had different requirements that were difficult for insurance companies to meet, he said.
“If we allow states to do it on their own, there will be friction,” Greer said. “This was created for a federal fix.”
Committee Chairman Rep. Joe Silesi (D-Montgomery) adjourned the session without a vote, but concluded by saying, “Congress is falling behind…and it's only going to get worse.”
“I'd like to see the federal government act, but sometimes the states need to move first and make the federal government wake up and do something, because there's a lot of haphazardness at the moment,” Silesi said. “The federal government is kidding.”
