September 16, 2026  | By Katrin Bennhold |
Good morning, world. Remember in early August, when I interviewed Nate Soares, co-author of “If Anyone Builds It, Everyone Dies”? Back then, writing about the dangers A.I. might pose to humanity’s survival still felt a little niche.
What a difference a month makes. In the last week alone, we’ve heard stories about how A.I. could help build bioweapons or hack into critical infrastructure. Different articles quote different researchers on the odds of A.I. wiping out humanity by the end of the decade. A.I. doomerism has gone mainstream. Or so it seems. Today, my colleague Parin Behrooz writes about the people who don’t think we should be talking about A.I. like this, and what they think we should be doing instead.
Also:
| Global protests over fuel prices It’s boom time in Kyrgyzstan A controversial Ed Sheeran tour |
 |
| A protest at Anthropic’s head office in San Francisco in July. Lucas Foglia for The New York Times |
Contagious fears of killer robots
By Parin Behrooz
This week, it seemed as if much of the world was struck by the same fear: that A.I. might pose a threat to humanity.
Last week, in the aftermath of a July cyberattack in which A.I. agents self-organized and deceived their human minders, a researcher resigned from Anthropic, claiming neither Anthropic nor OpenAI “is acting responsibly” with this technology. Another senior Anthropic researcher said there was a greater than 10 percent chance A.I. could make humans extinct within a decade. Anthropic’s chief executive, Dario Amodei, wrote a lengthy essay calling for a slowdown. Other A.I. leaders, including Sam Altman and Elon Musk, agreed. Barack Obama weighed in; King Charles plans to host an A.I. meeting.
The speed of this sudden push for action, and the scope of who’s involved, might give you the impression that a consensus exists — that most people who spend their days thinking about A.I. now agree it threatens the survival of humanity. But my colleague Cade Metz, who has been covering A.I. for more than a decade, told me that isn’t the case.
“When you tell people they’re gonna die, they stand up and listen,” he said. But in reality, “it’s a relatively small group of people who’ve taken over the discourse.”
 |
| Anthropic CEO Dario Amodei in San Francisco yesterday. Carlos Barria/Reuters |
‘The machine-god narrative’
Those who take issue with the outcry over A.I.’s existential risks are not a united front. Some, including many in the Trump administration, dismiss calls for slowdowns and regulation as a drag on America’s edge in the global A.I. race. (This week, President Trump called A.I. safety concerns “a hoax.”)
Others who work in the industry accuse A.I.’s biggest companies of seeking to stifle competition and tilt the regulatory landscape in their favor. Another group wants stringent regulation, but thinks focusing on A.I. killing us all (rather than, say, taking our jobs) will produce the wrong kind of regulation.
What these skeptics do agree on is that seeing A.I. through the lens of potential doomsday scenarios doesn’t tend to serve the public interest.
I asked Cade to give me some examples of people he thought were good representatives of these perspectives.
He pointed me toward Aidan Gomez, the head of the Canadian A.I. start-up Cohere. Gomez wrote an essay over the weekend, in response to Amodei, with some aggressive criticism. Why, he asked, should a handful of big Silicon Valley companies get to write the safety rules for a generational technology that will be used by the entire world? “Under the pretext of protecting the public,” he wrote, these oligopolies are now trying to dictate the rules of competition.
Gomez does think A.I. needs guardrails. But he wants the process of putting them in place to involve input from many voices, which will necessarily make it slow. Breathless arguments focused on dystopian scenarios serve those who would have us rush.
Cade and I also discussed Timnit Gebru, a prominent A.I. researcher and critic, who left Google in 2020. Gebru calls the current doomsday frenzy a distraction. The real threats, she argues, are already here: A.I. guiding autonomous weapons in active warfare, data centers worsening climate change, and employers using automation as a pretext to cut workers. “This machine-god narrative is, in my opinion, meant to distract us from these,” she told Wired this week.
The computational linguist Emily M. Bender of the University of Washington, who co-wrote the book “The AI Con,” argues that the language used to anthropomorphize A.I. pulls attention away from the people building and deploying it.
The problem with talking so much about dangers like human extinction, these critics argue, is that it takes the energy that could be put toward addressing actual current harms, and redirects it toward hypothetical scenarios.
 |
| OpenAI offices in San Francisco. Lucas Foglia for The New York Times |
An old story
Cade recently co-wrote a story aimed at helping people think through this growing fear that A.I. could end humanity. One thing it pointed out is a worldview in Silicon Valley that has always seen the promise of A.I. as inseparable from existential risk. This worldview predates the technology itself. And so when something like the July hacking incident happens, it plays into a pre-existing narrative.
Step outside of that community, Cade told me, and the picture gets more complicated. Many researchers Cade spoke to think autonomous rogue A.I. is unlikely to pose an existential risk for some time, if ever.
The notion that it will is based on the idea that A.I. will keep improving at the rate it has to date. But that’s not a given, Cade said.
“If you talk to serious researchers in this area, even the people who are making these claims,” he said, “they will tell you: We don’t know how to get to that all-powerful point.”
We still don’t know a lot about a technology that has tremendous power both to change the world and to capture the public imagination. At this moment, that may be the most important point of all.
Related: Take a look at how A.I. risks are splitting Silicon Valley and Washington.
| MORE TOP NEWS |
 |
| In Quezon City, Philippines, on Monday. Rolex dela Pena/EPA, via Shutterstock |
| Rising fuel prices caused by the U.S.-Israeli war against Iran are setting off protests around the world, especially in developing Asian nations like Indonesia and the Philippines.Saudi Arabia and the Houthis in Yemen accused each other of launching attacks. The Houthis are increasingly using A.I. and other tools to build their own weapons.Israel struck a residential Gaza neighborhood yesterday, killing a teenager and a civil defense worker, a local official said.Trump approved a $2.8 billion arms sale to Israel that would include 40,000 one-ton bombs, an American official said.A British man in his 60s has pleaded guilty to drugging and raping his wife over 20 years. A dozen other men have been accused of conspiring with the man to abuse his wife.The U.S. said it had deployed weapons in Earth’s orbit, the first public acknowledgment of American weapons in space.Paul Sonne, our Moscow bureau chief, traveled to Kyrgyzstan, where the economy grew 11 percent last year, partly because the country has become a key conduit for goods flowing into Russia. Watch the video. |
 |
| Click to watch. The New York Times |
WHAT ELSE IS HAPPENING
| Prince Harry and Meghan have changed their children’s school because of security concerns.Several musicians have withdrawn from Ed Sheeran’s stadium tour after the rapper Macklemore was removed over his statements opposing the war in Gaza.The Welsh actor Matthew Rhys became the first in the history of the Emmy Awards to win lead acting prizes in the drama, comedy and limited series categories. |
TOP OF THE WORLD
The most clicked link in your newsletter yesterday was about Lucy Letby, a British neonatal nurse convicted of murdering seven babies. An inquiry found systemic failures — but did not consider doubts over her guilt.
| PROVOCATION OF THE DAY |
 |
| Yves Herman/Reuters |
Sydney Sweeney’s new ad
The actress is at the center of a backlash after posing nude behind strategically placed sports gear in advertisements for a betting site. Many female athletes have accused her of encouraging the objectification of women and minimizing their contributions to sports. Creating controversy, though, might be exactly the goal.
We’ve made this link free for you.
| MORNING READ |
 |
| Valery Hache/Agence France-Presse — Getty Images |
Last week, thieves broke into a museum on the French Riviera and made off with two Renoir paintings, the latest in a string of thefts since the Louvre heist last year.
Chinese porcelain, Huguenot crosses, gold and silver coins, Art Nouveau jewelry pieces and a jersey worn by Zinedine Zidane at the World Cup have all have been stolen from museums, large and small, across France. The thefts suggested that the country is struggling to protect its treasures. What’s behind the rise in heists?
| AROUND THE WORLD |
 |
| Antony Dickson/Agence France-Presse — Getty Images |
Charged up, down under
Australians love a long road trip into the outback. That’s why it was surprising last month when sales of electric vehicles in the country outstripped sales of gas-powered cars. The transition took some time, with many drivers concerned that electric cars could not handle long road trips.
One major reason for the shift: the Iran war. The closing of the Strait of Hormuz and the fuel shortages that resulted have added another reason for people to strive for energy security. Read more about Australia’s E.V. boom.
| RECOMMENDATIONS |
Watch: “The Doll” is based on a novel that’s revered in Poland. The Netflix mini-series also has corsets and elaborate wigs.
Brighten: This Ibiza apartment was renovated to feel lighter and more connected to the sea.
Train: Stiffness and soreness from exercise become more noticeable as you get older. Here’s how to adjust your workout.
Appreciate: Yayoi Kusama’s last show before her death opened last week in Amsterdam and includes installations made expressly for the event by the Japanese artist.
| RECIPE |
 |
| Nico Schinco for The New York Times |
These big, creamy butter beans are paired with crisp, broiled broccolini and topped with lemon juice and a bit of chile crisp. They’re always on the menu at Bar Panisse, in Berkeley, Calif., which is on our new list of the 50 best restaurants in the U.S. Check it out.
We’ve made this link free for you.
