Hinton first laid out his concerns in a New York Times interview, saying he partially regrets the work he has done to bring A.I. to its current state. He’s worried about Google and Microsoft racing too vigorously to develop A.I. technologies, about propagandists flooding the internet with deepfakes—and about A.I. replacing human workers in “drudge work” roles. 

The psychologist turned computer scientist—who has previously been overly bullish on A.I.’s capacity to take over professions such as radiology—described the jobs threat as something that “might” happen. But is it really still a hypothetical? 

Not according to IBM CEO Arvind Krishna, who just announced a slowdown or suspension of hiring for roles that might soon be replaced by A.I. Krishna explicitly referenced back-office positions such as those in human resources, saying he could “easily see” around 7,800 roles “getting replaced by A.I. and automation over a five-year period.”

So much for the “copilot” rhetoric that is commonplace when promoting A.I. these days. Indeed, less than two weeks ago, Fortune carried a commentary piece from a prominent tech CEO who said his company’s human resources team had experienced “firsthand” how A.I. was tackling “the kind of tasks most people find repetitive, which frees up employees to take on higher-value work.”

That CEO was—you guessed it—Arvind Krishna. From his piece:

“With the help of IBM’s own A.I., we’re beginning to automate key steps of what was a very manual process. Piloting this technology within IBM itself, we have been able to shift from 700 professionals doing a relatively manual type of HR-related work, to less than 50. That’s freed up a very significant number of people to spend more time providing important talent-related services, such as career guidance and support for managers, which requires thought and creativity, rather than doing routine paperwork.”

Turns out some of those workers might find themselves rather more comprehensively freed up in the near future. It really is all happening so fast.

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. 

NEWSWORTHY

Samsung bans employees from using chatbots for work. In late March, staff at Samsung reportedly uploaded internal source code to ChatGPT, and now the company is concerned about security risks that come with using generative A.I. in the workplace. So concerned that employees who use it could face disciplinary action, including being fired. Bloomberg reports that Samsung recognized A.I. programs could save time on tasks, and told employees it would offer its own tools for translation and document summary. The move comes after other major companies have restricted or banned workers from using chatbots in recent months, like Amazon warning employees earlier this year not to upload confidential information to such services.

FAA faces lawsuit after Starship explosion. Environmental groups have sued the Federal Aviation Administration over SpaceX’s rocket explosion last month, saying the agency failed to fully consider the environmental impacts of the Starship program near Boca Chica Beach, Texas. The groups asked the court to revoke the five-year license the FAA granted to SpaceX. This comes after the U.S. Fish and Wildlife Service reported concrete masses, stainless steel sheets, and metal thousands of feet from the launchpad last week. The Associated Press reports that CEO Elon Musk said SpaceX would be ready for another Starship launch in the next six to eight weeks with FAA approval. 

YouTube positions itself as Gen Z’s go-to video platform. It’s no secret that Gen Z—basically anyone between the ages of 11 and 26—is very fond of TikTok. But YouTube wants brand advertisers to know that youngsters love its video site, too. At the company’s NewFronts event Monday, YouTube said it can achieve “awareness” and “reach” through campaigns on YouTube Shorts, its short-form video product. The event is the first in a string of NewFronts events from the digital platforms, including Snap, Meta, and TikTok, and comes after many of the companies delivered quarterly results that showed resilience in ad spending.

SIGNIFICANT FIGURES

24%

The growth in rides that Uber saw for the quarter ended in March compared with the same time last year. Its quarterly results, released Tuesday morning, surpassed Wall Street estimates and sent its shares up roughly 10%. Uber’s main U.S. rival, Lyft, which is dealing with layoffs of hundreds of employees, reports its results on Thursday.

IN CASE YOU MISSED IT

Elon Musk flies private jets more than any other billionaire—releasing over 2,000 tons of carbon emissions in his wake, by Chloe Berger

New Venmo feature lets users transfer crypto to outside wallets—and to each other, by Ben Weiss

Is $100,000 around the corner? Bitcoin posts fourth straight month of gains for first time since 2021, by Marco Quiroz-Gutierrez

The full Super Mario Bros. and Avatar 2 movies were reportedly uploaded to Twitter and viewed by millions, by Chris Morris

Tesla raises the price on Model 3 and Model Y after a series of cuts, by Chris Morris

BEFORE YOU GO

Turning thoughts into words. A system developed by researchers at the University of Texas might help people who are conscious but unable to speak (such as those who have had strokes) communicate with text. It works by measuring brain activity after people listen to hours of podcasts in a scanner. Later, their imagined telling of a story helps the machine generate corresponding text from brain activity. It cannot provide the thoughts word for word. However, researchers say the machine can provide text that closely matches the intended meanings of the original words.

Alex Huth, an assistant professor of neuroscience who helped lead the study published in Nature Neuroscience, said it’s a leap forward, and especially attractive since it involves a noninvasive medical procedure. “We’re getting the model to decode continuous language for extended periods of time with complicated ideas,” Huth said.