(The Center Square) – The “nonhuman status” of artificial intelligence chatbots would have to be disclosed to humans – and those same chatbots would have to be rigged against suicide-encouraging messaging – under a bill that passed the Pennsylvania Senate Tuesday afternoon.
Prime sponsor Sen. Tracy Pennycuick of Lower Salford Township in Montgomery County said AI is creating tremendous opportunities and tremendous risks. Specifically, there have been a growing number of cases in which children have turned to “AI chatbots” for emotional support and a “sense of connection,” and in such cases the robot messaging can seem very real, Pennycuick said.
Both she and Democratic Sen. Nick Miller of Allentown – the majority and minority chairs of the Senate Communications and Technology Committee – called the bill a common sense step toward protection.
The bill passed the Senate in a 49-1 vote, with Republican Sen. Doug Mastriano of Franklin County the only ‘no’ vote. It now goes to the House for consideration.
The vote came a little more than three months after President Donald Trump’s Dec. 11 executive order setting up the possibility of federal lawsuits against states that create AI laws that conflict with federal policy.
However, child-protection state laws were made exempt in that order.
Another member of the state Senate committee, Democratic Sen. Lisa Boscola of Northampton County, said state lawmakers are pushing ahead with work on AI bills without much thought of federal positions.
“We are just trying to listen to our constituents,” Boscola said. “We cannot wait for Congress. They have no ability to act sometimes.”
Republican Sen. Devlin Robinson of Allegheny County, lead sponsor of a bill to ban cell phone use by students during the school day, said of the federal AI positioning, “I don’t worry about that whenever I go to vote.”
Among other things, the bill that passed the Senate Tuesday requires that when the chatbot operator knows the user is a child, it should provide notices at least every three hours reminding the user to take a break and “that the AI companion is artificially generated and not human.”
It also directs the creation of measures to prevent the AI companion “from producing visual material of sexually explicit conduct or directly instructing the minor to engage in sexually explicit conduct.”
Enforcement would be provided by the state Attorney General’s office, with the potential for fines of up to $10,000 per violation.