cognitive surrender – Across TikTok and Substack, people are turning decision-making chatbots into a default option for choices that used to be uncomfortable and personal—what to eat, what to wear, how to message friends, and even how to leave toxic relationships. Researchers warn
On some mornings, it’s not a big life choice. It’s a plain one: what to eat, what to wear, how to respond. The twist is that more people are handing those moments to a chatbot—then returning to it again and again for reassurance.
Over time, researchers say, that pattern can quietly change what people get practice at. The less often they make hard decisions on their own, the more their confidence in their own judgment can shrink. In more extreme cases, AI researchers warn, the outsourcing could extend beyond actions to beliefs.
Cornelia C. Walther. a Wharton senior fellow and pro-social AI researcher. put it bluntly: “We want to believe we’re becoming more powerful thanks to our [AI] tools. ” she said. “But in fact. we’re giving away ever more power.” Walther added. “We are only half a step away from reliance and maybe in a not-too-far future full-blown addiction.”.
For Carolyn Yoo, the attraction wasn’t theoretical. A former software engineer who worked in tech for nine years until mid-2024, she described wrestling for months with whether to leave tech. Then she turned to Anthropic’s chatbot, Claude.
Yoo spent up to two or three hours a day asking Claude to map out different versions of her future. “I kept seeking AI’s reassurance to confirm whether this was a good life choice,” she said. She described a loop that kept pulling her back: “Every single day. whether I was talking to a new person who made me feel a bit unsure. I would go back to it again.”.
She wasn’t alone in leaning on chatbots for relationship decisions. In a Substack post titled “I’ve Outsourced My Judgement to AI. So Has Everyone Else,” financial writer and comedian Dominic Frisby described how he uploaded his entire WhatsApp exchange with an ex-partner for advice.
He wrote that he was “stuck in a toxic relationship I couldn’t seem to break out of, even after we separated.” At one point, he wrote, “I thought I was going mad.” The chatbot’s explanation of their relationship and personality types gave him clarity on how to move on.
Some apps are now trying to monetize indecision directly. Moot, launched earlier this year, offers an AI-powered decision-making experience in which users ask questions, then different AI personas debate them and vote on the best path forward.
The warning coming from researchers is that the practice itself—outsourcing evaluation until the “answer” feels final—may weaken the mental habits people rely on.
Wharton researchers have dubbed the phenomenon “cognitive surrender. ” describing a shift from treating AI as a tool to treating it as an authority. In that framing, users accept AI’s answers rather than actively evaluating them. Steven Shaw. a postdoctoral researcher in marketing at Wharton. said people could become “passive followers of unthought thoughts. ” adopting ideas generated by AI without fully processing them.
Other researchers describe the same drift in different terms. John Nosta. founder of the innovation think tank NostaLab. has warned that heavy reliance on AI can create a “cognitive codependent relationship. ” where people may be more productive in the short term while losing the underlying skills the technology is replacing.
Vivienne Ming. a theoretical neuroscientist and founder of Socos Labs. described the process as a classic case of “use it or lose it.” By repeatedly outsourcing mental effort to AI. she worries people will engage less often in the kinds of thinking that strengthen memory. attention. learning. and decision-making.
Anat Perry. a Helen Putnam Fellow at Harvard University. focused on feedback—what happens when an AI system is built to please. “When AI systems are optimized to please. they erode the very feedback loops through which we learn to navigate the social world. ” she said. Over time. Perry added. that could recalibrate what people expect feedback to feel like. making honest human responses feel “unnecessarily harsh by comparison.”.
Still, not every user frames the switch as surrender. Some argue that outsourcing parts of judgment can prevent mistakes they’d regret.
Frisby wrote on Substack: “I’m done with bad decisions. I’ve made enough wrong decisions for one life. I’m 56 now. I just want to make optimum choices and have a really good next three or four decades, or however long I’ve got left.”
The pull is easy to understand, too. The article describes why AI is hard to resist: it doesn’t get tired. it doesn’t get impatient. and it doesn’t feel uncomfortable. It responds instantly and can mirror reassurance back in seconds—often making it an appealing substitute when people don’t want to hear uncertainty from another person.
Joanna Stern. chief technology analyst at NBC and author of “I Am Not a Robot. ” described that unsettling relationship after spending a year using AI throughout her work and personal life. “The most unsettling part was seeing the relationship you can have with a computer. ” Stern said in a recent interview on journalist Kara Swisher’s podcast. She also said: “Even for people that seem well-adjusted, you can get so, so sucked into it.”.
For Yoo, stepping back became less about rejecting AI and more about rebuilding the ability to live with not knowing. She started relying more on notebooks, meditation, writing groups, and conversations with other people.
“The whole purpose is to sit in our own uncertainty and come to our decisions ourselves,” Yoo said. “That’s the only way you can really feel integrity in the choices you make.”
Between the reassurance people seek and the skills researchers fear are being traded away, the question now isn’t whether AI can advise. It’s what happens to the decision-maker when the advice becomes the default.
AI decision making cognitive surrender chatbots Anthropic Claude Moot app cognitive codependent relationship pro-social AI Wharton research toxic relationship advice human judgment
So basically TikTok ruined everyone’s brain, got it.
I mean… people been asking other people for reassurance forever. Now it’s just a chatbot. Does it really “fade skills” or is this just fear mongering?
Wait, is this saying AI can vote on your life like literally? Because that sounds like some sci-fi thing, like elections or something. I thought it was just helping you decide what to wear 😵
I get the idea, but also I’ve been using chatbots to text my boss and it saves time. Like if you’re too anxious to decide, you’ll probably just do it again anyway. Still though, “beliefs” getting outsourced?? That’s wild. Next they’ll say your personality fades because you asked ChatGPT what to eat on a Tuesday.