AI tools continue to automate more parts of our lives — and in turn, we become more reliant on them. Students use ChatGPT to do their homework, programmers ask Claude to churn out code, and now regular folks are getting a taste of AI agents with the launch of Meta’s cutesy personal assistant Muse.

If the trend continues, where does it lead? Some experts who spoke to CBS News worry that we’re on a path to slowly ceding more control to AI models until we become empty, waiting vessels for a bot’s decisions. There’s even a name for this phenomenon: “gradual disempowerment.” Its threat goes beyond job displacement, potentially affecting human cognition and agency itself.

“I put gradual disempowerment on the very top of my list,” Jay Chooi, the founder and CEO of Robocurve, a public benefit corporation that evaluates the role of robots in the workplace, told CBS. “A lot of people feel very powerless, because it requires a lot of effort and a whole-of-society approach to solve it.”

Jan Kulveit, a researcher at Charles University’s Center for Theoretical Studies and coauthor of the paper that coined the term gradual disempowerment, argued that it “does not need to emerge from a deliberate scheme or powergrab by AI systems.” 

“In the short-term, it is being incentivized by the perceived value that AI systems can bring to economic, cultural and state functions,” he wrote in the paper.

This is something many of us already picking up on. The people in our lives who ask an AI to do and plan everything are mocked as “meat proxies”; some are so thoughtless as to let an AI lay out a hiking trip that gets them stranded on a notoriously perilous mountain.

A more extreme imagining of this scenario portrays AI and robots taking over our jobs as they become a cheaper and more efficient alternative to human labor, driving companies that cling on to their human workforce out of business. It doesn’t need to be that drastic, though, to be harmful. Experts are already seeing signs of “cognitive surrender” among AI users — students especially — in which they’re effectively allowing an AI tool to supplant their own thinking. Research has linked reliance on AI tools to diminished critical thinking skills and demonstrated that AI leads to lower brain activity during cognitive tasks like essay writing.

On a societal scale, once we get trapped in this loop, it gets harder to get out.

“If people aren’t actually practicing the skills that the agents are doing for them, they lose the ability to have that skill, and so now you become reliant,” Margaret Mitchell, an AI research scientist at Hugging Face, told CBS. 

“What happens when something goes wrong?” she continued. “There’s no human you can point to who can fix it. So now you’ve become fully dependent on these systems to fix themselves, and what if that’s not possible?”

The decisions that AIs are being allowed to influence — or even make — are already incredibly consequential. An AI targeting system was reportedly involved in the decision to bomb a girl’s school in Iran, killing over 150 people, the vast majority of them children. President Donald Trump was reportedly encouraged to invade Venezuela and capture its president Nicolás Maduro by a conversation with Elon Musk’s Grok.

“A lot of people are talking about rogue AI hacking into data centers and taking over the world,” Chooi said, “but it is not in the mainstream consciousness that like AI could literally take over the world just by continuing to do what it’s supposed to do, which is automating everything.”