PCMag editors select and review products independently. If you buy through affiliate links, we may earn commissions, which help support our testing.

Experts: 5% Chance AI Kills Us All

A Berkeley, California, think tank asks AI experts to weigh in on when artificial intelligence will be able to perform tasks, such as producing a top pop song or rising up and destroying us.

 & Michael Kan Principal Reporter

Our team tests, rates, and reviews more than 1,500 products each year to help you make better buying decisions and get more from technology.

Our Expert
LOOK INSIDE PC LABS HOW WE TEST
65 EXPERTS
43 YEARS
41,500+ REVIEWS
(Photo by Yamada Hitoshi/Gamma-Rapho via Getty Images)

The latest advancements in AI have caused many to wonder if the technology will eventually doom us all. But a survey of hundreds of AI experts predicts the chances of artificial intelligence destroying humanity is actually quite small, at about 5%. 

The findings come from AI Impacts, a think tank based in Berkeley, California, that reached out to over 20,000 experts who had published papers in the past year. It received input from 2,778 researchers, who were asked for their predictions on the future of AI development. In addition, 1,321 researchers were asked for the probability that AI would lead to the extinction or severe disempowerment of the human race.

The experts gave “a median 5% chance” that AI would cause human extinction. “Mean responses indicated an even higher risk, suggesting a nearly one-in-six (16%) chance of catastrophic outcomes—the same odds as dying in a game of Russian roulette,” the think tank says

(Credit: AI Impacts)

Although a majority of the respondents rated the extinction risk low, AI Impacts notes: “One in ten respondents put at least a 25% chance on extremely bad outcomes like human extinction, with 1% of respondents thinking there was at least a 75% chance of such outcomes.”

How the experts envision AI causing humanity’s extinction is less clear. But the survey also polled the researchers on the various ways artificial intelligence could harm society. A large number of participants flagged AI producing false information, manipulating public opinion, and helping authoritarian regimes control society as key concerns. 

(Credit: AI Impacts)

A large block of researchers also said they worry about AI systems being harnessed to create powerful tools such as engineered viruses, accidentally causing a catastrophe, and accelerating income inequality as potential threats. 

The other noteworthy part of the survey is how it asked the AI community to predict when the technology could perform tasks, such as producing a top 40 pop song, writing a best-selling book, and even folding laundry. The results show that many experts predict AI has a 50% chance of achieving such milestones within the next decade, if not sooner.

(Credit: AI Impacts)

“Experts now see a 50% chance of AI models outperforming humans in every task by 2047, assuming no major disruption to scientific activity — just over 20 years from now,” AI Impacts adds. “Notably, this estimated date is 13 years sooner than the date arrived at by a similar survey in 2022, in which respondents estimated such AI would not arrive until 2060.”

About Our Expert

Michael Kan

Michael Kan

Principal Reporter

My Experience

I've been a journalist for over 15 years. I got my start as a schools and cities reporter in Kansas City and joined PCMag in 2017, where I cover satellite internet services, cybersecurity, PC hardware, and more. I'm currently based in San Francisco, but previously spent over five years in China, covering the country's technology sector.

Since 2020, I've covered the launch and explosive growth of SpaceX's Starlink satellite internet service, writing 600+ stories on availability and feature launches, but also the regulatory battles over the expansion of satellite constellations, fights with rival providers like AST SpaceMobile and Amazon, and the effort to expand into satellite-based mobile service. I've combed through FCC filings for the latest news and driven to remote corners of California to test Starlink's cellular service.

I also cover cyber threats, from ransomware gangs to the emergence of AI-based malware. In 2024 and 2025, the FTC forced Avast to pay consumers $16.5 million for secretly harvesting and selling their personal information to third-party clients, as revealed in my joint investigation with Motherboard.

I also cover the PC graphics card market. Pandemic-era shortages led me to camp out in front of a Best Buy to get an RTX 3000. I'm now following how the AI-driven memory shortage is impacting the entire consumer electronics market. I'm always eager to learn more, so please jump in the comments with feedback and send me tips.

The Best Tech I've Had:

  • My first video game console: a Nintendo Famicom
  • I loved my Sega Saturn despite PlayStation's popularity.
  • The iPod Video I received as a gift in college
  • Xbox 360 FTW
  • The Galaxy Nexus was the first smartphone I was proud to own.
  • The PC desktop I built in 2013, which still works to this day.

Read full bio