Actor Joseph Gordon-Levitt was in Utah for the Sundance Film Festival when he made a quick stop at the state Capitol to show support for the Artificial Intelligence Transparency Act.
The actor, known for his roles in “Inception” and “10 Things I Hate About You,” spoke in support of HB286 at a House Economic Development and Labor Services Committee hearing on Tuesday. A bill sponsored by Rep. Doug Feifia (R-Herriman) would require AI companies to publicly disclose child safety plans and risk assessments of their AI models.
Gordon-Levitt noted that he believes AI is a big issue that will ultimately impact everyone’s lives.
“It’s already very powerful, but it’s getting more and more powerful all the time, and many of its impacts will be major,” he said, noting that it will help increase productivity and advance science and medicine.
“But like any powerful technology, it can be helpful, it can be harmful, and it can be both. It all depends on how we use it, right?”
The bill has received national attention. Fiefia said this is because other states have enacted AI transparency laws, but Utah is the first to include child safety.
Here in Utah, it is rare for key figures to appear at committee hearings. Fifia worked with a number of national groups in the development of HB286, and these groups were what connected him to Gordon-Levitt.
What Joseph Gordon-Levitt said about the need for AI regulation
Gordon-Levitt spoke for about four minutes before leaving the hearing. As he left the building, multiple state lawmakers introduced themselves to him, some posing for photos with the actor.
He introduced himself as someone who works in television and film and has started a tech startup. He added that he is directing a film about AI for Netflix. Gordon-Levitt has recently been outspoken on AI issues, attending a number of events including the AI Summit in Utah last December.
“The question is: What are the principles, the morals, that guide the development and design of this technology?” he says. “And I’ll tell you from what I’ve learned, there’s only one principle in place right now, and that’s to make money.”
Gordon-Levitt pointed out that society works best when there is a balance between market incentives and the public interest. He said there is currently no balance in the AI industry.
“It’s out of balance because there’s effectively no law governing it,” he says. “There are more laws governing how sandwiches are made and sold than there are laws governing this incredibly powerful, innovative technology that will change everything about our lives.”
Gordon-Levitt revealed that she has three children, ages 10, 8, and 3.
“I worry that our children will grow up in a future dominated by unethical AI businesses that have proven time and time again to be incapable of prioritizing children’s well-being,” he said.
What will Utah’s AI bill do?
Utah has been at the forefront of developing and regulating AI. In 2023, the Legislature established a first-of-its-kind AI Policy Lab to guide innovation and help develop consumer protection reforms.
“Some of today’s AI systems are very powerful. If they fail or are misused, the damage will be more than small. It can spread quickly and affect many people at once,” Fifia said.
Fifia, who previously worked at Google, authored the bill to protect children who are most harmed by AI failures.
Sponsors said the bill, if passed, would do four “simple things.”
Require businesses to post public safety and child protection plans on their websites. Require companies to be honest about the risks of AI models. Require companies to report incidents when they occur. Protect whistleblowers so that engineers and other employees can speak openly about safety issues without fear of retaliation.
The bill would impose a civil penalty of $1 million for a first violation and a $3 million penalty for subsequent violations.
On Tuesday, Fiefia filed a replacement bill that could transfer funds from those fines to a fund and cover the bill itself.
Critics of the bill say it could stifle progress in AI development in Utah and run counter to the Trump administration’s AI goals.
Fiefia denies these claims, saying: “There are no content mandates, no government pre-approvals, and no algorithmic micromanagement. We don’t touch development, which means we don’t stifle innovation.”
He is working on a second bill that would address how AI chatbots are used by minors. The matter has not yet been made public.
“Both bills actually come from the same place: AI is already shaping children’s lives, and the rules just haven’t kept up,” Fifia told the Deseret News.
During the hearing, Abundance Institute Chief of Staff Mae Kennedy opposed the bill. She said her organization supports transparency and child protection, but said the bill lacks important details.
“There are some aspects of the bill that lack specificity, particularly the definitions, that do not help create the desired effect we are seeking,” Kennedy said. “If we are going to regulate AI, new technologies like this require some specificity and clarity.”
Joseph Gordon-Levitt’s AI advocacy
In December, Gordon-Levitt attended the Utah AI Summit with Utah Governor Spencer Cox. He spoke at the event about how society should prevent artificial intelligence companions from replacing real human relationships.
As previously reported by the Deseret News, Gordon-Levitt said, “The foundation of every civilization is relationships.” “There’s more to relationships than a chatbot can do.”
In September, Gordon-Levitt posted a video to the New York Times saying, “Meta AI chatbots are dangerous to children.”
In 2025, he gave several other speeches on AI, primarily focusing on the need for AI regulation and how generative AI models utilize the work of many artists, as previously reported by the Deseret News.
Speaking at the United Nations’ Internet Governance Forum in June, his short speech focused on AI regulation and ended with a simple principle:
“Your digital self should belong to you. The data that humans produce – our writing, our voices, our connections, our experiences, our ideas – should belong to us. And the economic value generated from this data should be shared with the humans who generate it,” he said.
