6 scary predictions for artificial intelligence in 2026

When OpenAI When it announced a “code red” this month to refocus its teams on competing with Google, I couldn’t help but think of December three years ago when the companies’ roles were reversed. It was Google that sounded the sirens to catch up with OpenAI. What followed the next month, in January 2023, were the first large-scale layoffs in Google’s history. “A difficult decision to prepare us for the future,” as the company described it at the time.

I wonder if the ChatGPT developer could make similar workforce cuts early next year. This speculation inspired me to come up with a whole bunch of predictions about what might come next year. Here’s a look at six of the ideas, fine-tuned with the real intelligence of WIRED colleagues.

Data center misinformation

Communities around the world are struggling to build data centers. In the United States, many activists organize themselves on social media using tools such as Facebook groups. The Chinese and Russian governments continue to exploit social media to spread disinformation disguised as real news and real opinions. A slowdown in data center development in the United States would be a boon for China and Russia, which are seeking to surpass the United States in industrial and military AI capabilities.

Austin Wang, a researcher at the nonprofit RAND Corporation who has studied Chinese-controlled propaganda farms, says there are no signs of troubling activity at the moment. “Many of the newly created anti-data center pages seem to be controlled by real American citizens so far,” says Wang.

But as anti-data center fervor grows, China and Russia could try to boost grassroots organizing. And the work is made easier thanks to artificial intelligence that can quickly create photos and videos to impress people on social media.

Robot demos are everywhere

In 2026, tech conferences, from the Consumer Electronics Show to Amazon’s hardware event, will likely be full of talk about AI-powered robots. Google and other big tech companies have spent years trying to train robots to handle household tasks through repeated practice. But now there’s a new round of hype. The kind of AI models used in services like ChatGPT and Gemini are being incorporated into robots in the hope that they will handle household chores, such as folding clothes, with less training and greater accuracy.

Last September, Google released a video of a robot that sorts trash, compost, and recycling in response to a user’s voice commands. When Google executives take the stage at the company’s upcoming I/O conference, I expect them to have the robot do tasks like, say, sticking a pizza into a kind of oven it’s never seen before and, while it’s cooking, retrieving a half-full can of Diet Coke from the back of a cluttered refrigerator.

Barak Turovsky, the recently departed chief AI officer at General Motors and a former AI leader at Google, says advances in robotics capabilities are possible because large language models can understand a dishwasher manual, learn how to operate a dishwasher by watching a video, and understand how to grab a particular part by decoding a drawing. “The next frontier for large language models is the physical world,” he says.