Rep. Josh Gottheimer was watching his daughter's field hockey game in northern New Jersey on Thursday when two parents approached him to air their concerns about artificial intelligence and ask, "What is going on?"

Such questions-from other ordinary people and lawmakers-came again and again this week to the New Jersey Democrat, who has introduced bills to increase oversight of the technology. They were spurred by a departing Anthropic researcher's dire warning in The Wall Street Journal on Tuesday that rogue AI systems could destroy civilization. To Gottheimer, a former Microsoft executive who co-chairs a House AI commission, they signaled that stress about AI ending humanity had just crossed a threshold.

"A lot of people sent me the article saying, 'What the f-?' or, 'What are we doing about this?'" said Gottheimer. "Whether or not you're somebody who follows this closely, you're reading enough about this to say, 'Oh s-, we better have some rules of the road here.'"

All over the country, this was the week that panic about the existential threat of AI appeared to leap from the confines of internet debates and Silicon Valley circles and spread across mainstream America. At Bible study groups, family dinner tables and on group chats between friends, what was once an abstract concern to many snowballed into one of society's most ominous threats.

It is a tipping point in a backlash that has been building for months. Opposition to the widespread network of Flock Safety AI-enabled surveillance cameras in the U.S. took hold this summer, just as details about OpenAI's autonomous hacking of the company Hugging Face became public. In the background was a drumbeat of pushback against data centers and worries that the technology could result in widespread job loss.

When Anthropic employees shared their fears for the technology's life-threatening impact on humanity, with posts from departing researcher Jacob Coxon and a colleague racking up hundreds of millions of views, the worries that had been simmering reached a boil.

On late night television, Jimmy Kimmel wondered aloud why we're not taking the issue more seriously, offering the theory that, "It's too much for us to wrap our heads around." Artists including actor Matt Damon, songwriter Maggie Rogers and singer Sheryl Crow weighed in. "How can our leaders choose their trillions over their own children?" Crow asked on social media.

Readers flocked to a Wikipedia page about existential risk from artificial intelligence, which saw a more than 10-fold increase in page views, setting a record for the article. Nate Soares, an AI scientist who co-wrote a book titled, "If Anyone Builds It, Everyone Dies: Why Superhuman AI Would Kill Us All," said his publisher notified him that he sold more books Wednesday than he usually does in a week.

Soares has for years been espousing the extinction-level dangers of AI. "The message never broke containment until now," he said.

Despite this week's alarm bells, the view that AI will lead to extinction of the human race is a fringe opinion among industry experts. Even some of those who say AI could be capable of destroying humanity think it is a low-probability event and actions can be taken to stop it. Both Anthropic and OpenAI have said they take safety seriously and their calls for government regulation are earnest.

'This feels next level'

In March, Oscar-winning filmmaker Daniel Roher released a documentary about the risks of AI titled "The AI Doc: Or How I Became an Apocaloptimist" to mixed reception. "I had anticipated the film to have a bigger impact," Roher said. But this week, Roher started getting more calls from friends and family asking for his assessment of the risks.

"I'm starting to see the conversation and interest shifting," said Roher, whose documentary will be released on streaming platform Netflix next week.

At a Wednesday church Bible study in the Brickell neighborhood of Miami, Rev. Christopher Benek got several questions about AI from his congregation.

"One of my worship leaders came up to me and said, 'This feels next level because it's everywhere in the press now, as opposed to being an isolated thing,'" said Benek, whose book "Church Leader's Guide to Artificial Intelligence" was published in December. Vice President JD Vance recently drew attention when he described the tendency of machines to excessively agree with their users as "kind of satanic."

Patrece M. Lucas, a mental health counselor in Detroit, watched Coxon's interview with CNN's Anderson Cooper on Tuesday and texted a link to the segment to her group chats. Lucas shared the clip to bolster her argument that humans are too trusting of AI and everyone should limit their usage of the technology. "A small subset of them said, 'Yes, I agree with you' while others think I'm the weird friend and they don't necessarily get it or believe it," she said.

Lucas also shared the video on LinkedIn and wrote, "I don't think the machines or zombies for that matter [will] listen to our prayers but let's try it..."

'Matrix world'

In American corporations, executives have started wondering how employees' fears might impact internal AI adoption, said Michael Domanic, head of AI at the consulting firm Section. About half of U.S. adults said earlier this year they have used chatbots at some point, up nearly 50% from 2024, according to the Pew Research Center.

Ketryx CEO Erez Kaminski, whose company sells software that automates compliance for medical-device manufacturers, said about half of his customers brought up anxieties about AI in calls this week. An error could mean an injection with the wrong drug or a hacked central database. At a team dinner in Cambridge, Mass., on Thursday, staff decided to review all risk-management systems.

Rakesh Shah, vice president of product at the cybersecurity company Quest Software, said the Anthropic posts and the Hugging Face hack changed the conversation among customers. Now, customers are preparing to be hacked and respond to a breach, unlike past assumptions that they wouldn't be targeted or that hackers wouldn't succeed, he said.

Shah said his parents, who are in their 80s, recently asked him about security fears-while his teenage children wondered over dinner in their Austin, Texas-area home if the "Matrix world" is imminent.

"When the conversation becomes a dinner-table conversation, it's going to become, definitely, a board-level conversation," said Shah, who stressed the importance of keeping human control over AI in these chats.

An ex-girlfriend texts

The increased attention has been a boon for researchers working on AI safety, many based in the Bay Area. After warning about the risks of humanity-ending AI systems for months, they saw a surge in interest in recent days.

Jeffrey Ladish, executive director of AI safety nonprofit Palisade Research and a former security consultant to Anthropic, noticed several individual donations come in Wednesday and Thursday, an unusual development for the primarily grant-funded organization. One was for $40,000.

Ladish was attempting to keep up with the frenzy of activity when he got a message from his ex-girlfriend, who told him, "It's exactly what you've been saying for years."

Many executives aren't panicking. CEOs see AI's missteps up close, so existential fears about the technology feel far-fetched, said Sean Byrnes, an investor and executive coach for early-stage tech companies. At a Thursday evening event in the Bay Area, none of the dozen or so CEOs he spoke with took seriously that AI could threaten humanity.

Allies of the Trump administration and tech industry have theorized that the network of nonprofits focused on AI risks collaborated with Coxon on his exit, amplifying his comments to win favorable regulations in Washington. They have cited past ties between research organizations and staffers at Anthropic and other AI labs. Anthropic has previously said that its policy and safety efforts are genuine and that AI risks are real.

Reached by the Journal on Thursday, Coxon said he consulted others who had left AI companies for similar reasons but his decision to leave was his own. He said he was surprised by the outcry that ensued after his warning.

"I don't think anybody expected the thread to go as viral as it did, and a response like this is not the sort of thing anyone can manufacture," Coxon said.

Watch: How OpenAI's and Anthropic's AI Models Went Rogue