Most Americans see A.I. as an existential threat. The worries are reasonable but they could make things worse

PETER PARKS/AFP via Getty Images

Most Americans are pretty terrified of artificial intelligence, according to a survey just released by Reuters and Ipsos. 

A whopping 61% of respondents thought A.I. could threaten humanity itself, with just 22% coming out against that proposition. Interestingly for those trying to predict the political response to A.I.’s sudden ascent—more on that later—the survey found Trump voters were notably (70% vs. 60%) more likely than Biden voters to perceive an A.I. threat to our species, and Evangelical Christians were markedly more likely than non-Evangelical Christians to “strongly agree” with that threat’s existence (32% vs. 24%).

These stats are really remarkable, especially given how recently generative A.I.’s appearance started sounding alarm bells. However, they are also unsurprising due to the extreme volume of those warnings, and the fact that many of those warnings come from the people building these systems: that Musk-and-Woz-signed open letter (“potentially catastrophic effects on society”); ex-Googler Geoffrey Hinton (“more urgent” than climate change); OpenAI CEO Sam Altman (the A.I. industry could “cause significant harm to the world”).

Of course, everyone’s freaking out—you can’t check the news these days without some expert yelling that the sky may soon fall.

I have mixed feelings about this. On the one hand, great! People are listening to experts again! I’ll take it! Of course, many less well-known experts may be justifiably annoyed that the public wasn’t listening to their warnings until recently, but that’s kind of understandable—the world is full of very pressing concerns, and serious A.I. warnings weren’t hitting the front pages or evening news a year ago.

On the other hand, there are two big problems with today’s heightened level of fear. 

Firstly, as umpteen people pointed out when that open letter was published, people really should be less concerned with A.I.’s theoretical existential threats—which serve to bolster the industry’s narrative that what they’re building is oh so powerful and therefore valuable—and more concerned with its existing effects on the spread of disinformation and the perpetuation of biases. 

Secondly, when everyone’s freaking out, there’s a whiff of moral panic in the air and a strong risk of bad laws emerging. That’s not to say lawmakers and regulators shouldn’t be addressing A.I.’s myriad issues at speed—they have no choice, and we need them to do so—but the paths they choose will have massive consequences for the industry and for society. 

As my colleague Jeremy Kahn noted in his must-read write-up of Altman’s Capitol Hill visit yesterday, the OpenAI CEO was, uh, open about the danger of new rules being designed in a way that leaves only big companies such as his being able to comply. Europe’s open-source community and some legal experts are also concerned that the EU’s almost-there A.I. Act would mandate risk assessments that small firms and projects just can’t handle, while potentially stifling some of A.I.’s less widely-appreciated positive use cases.

Do we want a future where only big companies are effectively allowed or able to train large A.I. models? Maybe we do (to keep as tight a lid on the tech as possible) and maybe we don’t (because stifling competition rarely ends well for the user). But whatever route we take, it needs to be well thought out and chosen on the basis of the facts. The majority of Americans seeing A.I. as an existential threat to civilization—which makes it highly political by definition—is unlikely to help that happen.

Want to send thoughts or suggestions to Data Sheet? Drop a line here.

David Meyer

Data Sheet’s daily news section was written and curated by Andrea Guzman. 


Tesla to try out advertising. During Tesla’s annual shareholder meeting Tuesday, CEO Elon Musk said that ads are worth testing out. The EV maker has relied on word-of-mouth to push sales for its vehicles for the last several years, avoiding traditional advertising. But don’t expect a commercial ad pitching the Cybertruck anytime soon. “I only just agreed to it so it’s not a fully formed strategy,” Musk said.

SVB’s former CEO grilled at Senate hearing. During a two-hour hearing, former SVB CEO Greg Becker told Senators that he partially blamed the media for its coverage of SVB and said that a herd mentality set off a bank run. Members of the Senate Banking Committee pushed for answers on executive mismanagement, including an explanation for why SVB didn't have a chief risk officer in the eight months before the bank run. Becker said they were searching for a qualified candidate and that his team made the best decisions with the facts available to them at the time. He also denied acting on nonpublic information when he cashed out some of his stock options in the weeks before SVB fell.

Teens can now hail an Uber. Ride-hailing app Uber is releasing a teen accounts feature that will let parents hail rides for those aged 13 to 17 and then monitor them through the app. The feature will open up to users in New York City, Atlanta, Houston, and other metros in the U.S. and Canada later this month and will launch in more cities in the coming weeks. Teens will give their driver a unique PIN number at the start of the ride and there will be in-app audio recording during the trip. Parents can also contact the driver and Uber’s support team as the ride takes place. Still, Uber CEO Dara Khosrowshahi addressed safety concerns, saying that the option is restricted to drivers who are experienced and highly rated.


170 billion 

—The amount of rupees ($2 billion USD) that India is putting toward a financial incentive plan to attract laptop, tablet, and other hardware makers following the success of Apple's assembly operations in the nation. 


Elon Musk warns shareholders next 12 months will be challenging for Tesla as U.S. banks ‘on their way to the cemetery’ pull car loan financing, by Christiaan Hetzner

Bosses are freaked out you’re working 2 jobs—and remote work and AI have let the cat out of the bag, by Tristan Bove

Nintendo sells 10 million copies of The Legend of Zelda: Tears of the Kingdom in 3 days, by Chris Morris

Michael Lewis had a front row seat to Sam Bankman-Fried’s downfall. He called the disgraced former FTX CEO ‘the ideal subject’, by Marco Quiroz-Gutierrez

Peloton’s former CFO reflects on what she learned from the company’s rapid rise and fall, and her new role at data-driven startup Alation, by Sheryl Estrada


Amazon’s Echo updates. Product refreshes including the next-generation Echo Show 5 and Echo Buds have arrived. But the main product Amazon boasted about is its latest addition to the line today with the Echo Pop, a front-facing directional speaker that comes in lavender, teal, black, and white for $39.99.

Running on the company’s AZ2 chip with eero, a mesh wifi network system, built-in, the device has 1,000 square feet of coverage. It also comes as Amazon says Alexa has seen rising use and is experimenting with making Alexa more conversational. The Verge reports that the company is trying to incorporate generative A.I. and large language models into the assistant, but it will likely be a while before it’s perfected. “There are certain things that you don’t want to have to happen,” ​​said Dave Limp, SVP of devices and services at Amazon. “If you ask to buy paper towels, we need to get the price right, and we need to ship it to the right address.”

This is the web version of Data Sheet, a daily newsletter on the business of tech. Sign up to get it delivered free to your inbox.

Read More

CEO DailyCFO DailyBroadsheetData SheetTerm Sheet