Dozens of bills in Albany target potential risks of AI

ALBANY — One bill would ban artificial intelligence chatbots in children’s toys. Another would regulate how chatbots can dispense legal and medical information. A third aims to stop artificial intelligence algorithms from discriminating against protected classes.

Democrats in the state Legislature are considering dozens of bills to protect consumers who use AI from potential harms. The legislation comes in the wake of the state passing the RAISE Act last year, which set guidelines for large-scale AI models, and as the federal government is beginning discussions on a national AI framework to regulate the growing industry.

State Sen. Kristen Gonzalez, D-Queens, chair of the Senate’s internet and Technology Committee, said she believes the state still needs to enact broader legislation on topics such as data privacy and guardrails for high-risk AI uses.

"What we're finding is that there is an urgency to deal ... with specific use cases in AI that are more and more being documented to prove harm," she said.

For Gonzalez, one of those urgent areas is regulating chatbots that children use. She is sponsoring a bill that would ban chatbot operators from offering features to children that encourage emotional attachment with the chatbot.

It would also ban them from offering specific responses, like discouraging young users from speaking to adults or encouraging them to engage in sexually explicit conduct.

Gonzalez pointed to instances where children and young adults harmed themselves or even killed themselves. In one case in Maine, an AI chatbot allegedly offered to help the teen write a suicide note.

Last fall, a survey by the digital safety nonprofit organization Common Sense Media found that 72% of teens have used AI companions at least once. More than half use them several times a month.

"We don't want a chatbot humanizing itself," she said. "We want to make sure that at every point a consumer knows that they're engaging with a machine, not a person."

Industry concerns

A number of Gonzalez’s bills have drawn pushback from the tech industry. Earlier this month, the Software Information Industry Association sent her a letter calling her bill limiting chatbot interactions with children overly broad, vague or technically infeasible.

Julie Samuels, the president and CEO of Tech: NYC, a nonprofit representing the city’s technology industry, said it would be easier for the state to legislate AI if the state had a shared framework for how to treat New Yorkers’ data.

Twenty other states have comprehensive data privacy rules, she said. "It means that every time we address one of these issues, we're kind of playing whack-a-mole, as opposed to starting from a shared understanding."

Samuels said the majority of the AI bills moving through the state legislature are not objectionable on their face. The potential problems lie in the details, and her members believe it is important that the state doesn’t create a business environment where tech founders and companies feel like they have to leave New York.

"We have to balance creating safe, efficient, effective guardrails, with the understanding that we are in a moment of significant technological innovation, and we want to be in the game," she said.

Possible risks

Gonzalez said she is not against artificial intelligence. But she does have significant concerns about potential harms and whether regulation can keep up with a rapidly changing industry.

"That's a tension I think about every day," she said. "I believe that there is a future that we can achieve where artificial intelligence and emerging technologies are used to the benefit of society, but that can only happen if we agree on when and how and why these tools are being used."

Earlier this month, state Attorney General Letitia James lent her support to two bills that would ban the use of surveillance pricing, which relies on AI-powered algorithms to offer personalized prices to consumers.

Assemb. Alex Bores, D-Manhattan, who sponsored the RAISE Act last year, is sponsoring a number of AI-related bills including one meant to eliminate deepfakes, which are AI-generated images or videos which make it seem like a person is saying or doing something they haven’t done.

"AI is only growing and only touching more. And so, you're going to see bills of how it intersects with every area," he said.

Bores' bill was included in Gov. Kathy Hochul’s budget proposal and Bores said he is hopeful it will be part of the final budget package.

Another bill would extend the provisions in the NY SAFE Kids ACT, which bans social media companies from offering personalized feeds based on user data to minors, to adults.

Bores said in his discussions with colleagues on how to approach regulating the industry, he tries to keep the focus on the effects AI can have if the state doesn’t act.

"I actually think a lot is just keeping the human elements in these conversations and making clear, like, ‘hey, someone's getting scammed, a kid is getting put down a rabbit hole of disinformation,’ " he said. "These are impacts you can understand."