Bay Area teen working to promote AI safety education for children

By Amanda Hari

Governor Gavin Newsom signed a number of bills into law establishing guardrails for how Artificial Intelligence can interact with children and vulnerable individuals.

But he vetoed one, Assembly Bill 1064, that would limit children from using most companion chatbots.

High school senior, Kaashvi Mittal, 17, can understand why the bill was written.

“Obviously, kids are the most vulnerable population in terms of their ideas, they form very easily,” Mittal said.

Although still a teen, Mittal has already spent time studying AI through a program at Stanford called ‘AI for All’.

“It was through that program that I learned a lot about the possibilities of AI from discovering drugs to actually having mobile robots that can do different things, and then after that I founded an organization called Together We AI to make AI education accessible to everyone,” explained Mittal.

She’s aware of the dangers that come with it. Back in April, a Southern California 16-year-old took his own life after having a conversation about it with ChatGPT.

“Just what you mentioned is actually one of my biggest concerns,” Mittal detailed. “I’ve heard a lot of stories circulating about either AI models that are engaging in inappropriate relationships with minors or encouraging minors to do things that harm themselves or others.”

Which is why she wants to educate young people, so they have the tools to navigate AI safely, reminding them it’s not always correct and that it’s not a real being.

She appreciates Governor Newsom stepping in and signing bills that continue to create safeguards for artificial intelligence.

“We’ve seen some truly horrific and tragic examples of young people harmed by unregulated tech, and we won’t stand by while companies continue without necessary limits and accountability,” said Newsom in a statement. “We can continue to lead in AI and technology, but we must do it responsibly, protecting our children every step of the way. Our children’s safety is not for sale.”

While he signed a number of bills, he did veto one, Assembly Bill 1064. That bill would prohibit children from using many companion chatbots, like ChatGPT, unless it could be proven that they aren’t “foreseeably capable” of performing harmful behaviors, like encouraging self-harm or disordered eating.

Mittal understands the decision and can see the drawbacks of the bill.

“The bill used very broad terminology, and I know one of his concerns was if it’s so broad this bill could restrict certain AI technology that is useful for minors, like AI learning systems for example,” said Mittal.

She says AI, and chatbots specifically, can benefit minors, but she hopes going forward that engineers will focus on creating ethically sound AI systems.

“The biggest thing is really about testing and making sure that before these AI models are rolled out for the public to use that they are comprehensively tested and that we can be confident that these AI models won’t be encouraging people to do harmful things or spreading wrong ideas,” Mittal stated.