Actor Joseph Gordon-Levitt said society should guard against allowing artificial intelligence companions to replace real human relationships during Utah’s 2025 AI Summit on Tuesday.
Echoing remarks made earlier in the day by Gov. Spencer Cox, who introduced his “pro-human” AI agenda for the state that included business, education and policy components, Gordon-Levitt told Margaret Busse, executive director of the Utah Commerce Department, that there are limits to what AI can do.
“The bedrock of any civilization is human relationships,” Gordon-Levitt said. “There’s more to a human relationship than what any chatbot can do.”
Gordon-Levitt, a tech entrepreneur and founder of the online creative platform HITRECORD, said he’s not anti-technology — he’s an optimist who believes AI can help with education, health care and creativity. But he’s deeply worried about AI “companions” aimed at teens and children.
Gordon-Levitt joined Cloudflare CEO Matthew Prince and NVIDIA co-founder Chris Malachowsky in speaking at the summit on Tuesday, where they all addressed the benefits and threats posed by the increasing use of AI.
Joseph Gordon-Levitt on kids and AI companions
Gordon-Levitt pointed to leaked internal documents from Meta outlining highly sexualized language its AI companion could use with young users designed to manufacture “synthetic intimacy to hook” young people and keep them engaged for advertising dollars.
“What’s a great way to get your attention?” he asked. “It’s to sort of form a relationship with the users. And that’s what these AI companions are completely designed to do.”
In September, Gordon-Levitt published an op-ed in The New York Times and posted a video to social media where he said that “Meta’s A.I. chatbot is dangerous for kids.”
If the business model is maximizing attention for ad revenue, he said in Utah on Tuesday, parents should be “pretty concerned” about putting that system in front of their kids. By contrast, he said, there’s a world where AI tools in schools are built to help children learn and are “optimized for their well-being, not for business interests.”
Gordon-Levitt said he doesn’t support total bans — comparing AI companions to alcohol or gambling — but said society needs to name them honestly as “an indulgence and a vice and a trick,” not a substitute for friendship.
“If a person’s idea of a relationship is formed through interaction with this chatbot,” he said, “we’re headed for a civilization of people lacking empathy, lacking perspective, lacking the ability to really have a human relationship.”
Cloudflare CEO Matthew Prince: AI is about to rewrite the internet’s business model
Earlier in the day, Cox sat down with Cloudflare CEO Matthew Prince to explore how AI could reset the incentives that have shaped the modern internet.
“I can be anti-social media, but still pro-internet,” Prince said, describing the last 27 years as a “miracle” built on a simple model: create content, drive traffic, then turn that traffic into money through ads, subscriptions or app sales.
For social media, he said, “the key was, how do you generate more traffic?” That led platforms to favor content that “makes people either angry, generates cortisol, or gets them kind of excited,” driving what he called a world of “rage-bait.”
AI, Prince argued, is a new platform shift — like the web, then social, then mobile — that will change that model. Instead of sending users to “10 blue links,” AI agents will simply answer questions.
“That’s a different business model,” he said. “Inherently, that means the business model of the internet is going to fundamentally change.”
Prince, who owns the Park Record newspaper in Park City, said local news illustrates AI’s potential upside. After the paper and local radio station KPCW mailed a nonpartisan voter guide to residents, turnout in Park City rose from about 22% to 57% in an off-year election, he said.
“It is one of the most impactful things that you can do,” he told the audience, urging young people to consider starting or strengthening local news outlets — both for civic reasons and because “that might be the information that’s the most valuable to AI companies” in the future.
Prince also warned against a future with only “five AI companies,” accusing Google of using its dominant search “bot” to see “five times more of the internet than any other AI company,” including behind paywalls, and calling for regulations that would force Google to separate search crawling from AI training.
“Google should have to play by the same rules as everyone else,” he said, arguing that true competition is key to a pro-human AI future where small businesses and local media can thrive.
NVIDIA co-founder Chris Malachowsky: Utah as a model for AI literacy
In another conversation, Jefferson Moss, executive director of the Governor’s Office of Economic Opportunity, interviewed NVIDIA co-founder Chris A. Malachowsky about how AI is reshaping the economy and workforce.
Malachowsky shared NVIDIA’s origin story: three chip designers who saw an opportunity in the early 1990s gaming market to “build a game console on a chip” and create graphics hardware with “unlimited headroom” — enough parallel computing power that would eventually become ideal for AI.
“We got to AI on the backs of kids playing games,” he said with a laugh, noting that GPUs built for ever-better video games turned out to be “the most prolific and available parallel processor” for modern AI workloads.
Malachowsky praised Utah’s “bias for action,” pointing to the state’s AI partnerships with NVIDIA and its investment in an AI-ready workforce as a model he now uses on a 50-state tour encouraging similar efforts.
He said AI literacy should start early — not by putting kindergarteners in front of chatbots, but by teaching them how the systems work so “when they get older, and they’re using these tools, they know what they are, and they know what they’re not.”
Looking ahead, he said states can use AI to build “digital twins” of critical infrastructure, like Florida’s model of Jacksonville that simulates buildings, utilities and storm behavior to plan disaster response. Those kinds of investments, he argued, can improve resilience, public safety and long-term economic security.
“The key is to keep the human” in the loop, he said. “Teach the human how to be valuable, how to leverage tools and not be leveraged by them.”
