
Over half of Americans fear losing both their jobs and their independent thinking to AI, survey finds
Quick Answer
A survey by Anthropic reveals that 64% of nearly 52,000 Americans fear job losses due to AI, while 56% worry about losing independent thinking.
Quick Take
Interestingly, daily AI users show less concern, yet many reject AI in their workplaces for tasks they believe it can perform.
Key Points
- 64% of respondents fear job losses to AI.
- 56% worry about losing independent thinking due to AI.
- Daily AI users are less concerned about these issues.
- Many reject AI for tasks they believe it can handle.
- Survey included nearly 52,000 Americans.
📖 Reader Mode
~3 min readA major Anthropic survey of nearly 52,000 Americans shows widespread anxiety about AI-driven job loss and cognitive dependency.
Anthropic has published the results of its first representative survey series, Anthropic Public Record. The online poll, conducted by YouGov between November and December 2025, covers 51,993 Americans aged 16 and older across all 50 states, the District of Columbia, and Puerto Rico. According to Anthropic, it's the first survey aimed at the public rather than just users of its chatbot Claude.
Curing disease tops the wish list, job loss dominates the fears
When asked to name their three biggest hopes for AI, 48 percent of respondents picked curing diseases like cancer and Alzheimer's. Helping people with disabilities came in second at 36 percent. Technological progress and making everyday life easier tied at 23 percent each. The least popular hope was AI as therapist and cure for loneliness, a use case that's often discussed as a major opportunity but has already caused real harm.

The fears are dominated by job loss. 64 percent of Americans worry AI will cost them jobs. That concern grows with education level and ranks first in every single state, Anthropic says. Fear of cognitive dependency, meaning the loss of the ability to think for yourself, comes in second at 56 percent. Misinformation takes third at 52 percent. Only 15 percent trust AI companies to make the right calls about how the technology gets built and deployed.

Each of these fears has roots in earlier technologies, Anthropic notes, like automation, smartphones, and social media. In general, Americans worry far more about humans misusing AI than about the "AI goes rogue" scenario that parts of the AI industry love to debate.
Cognitive dependency: widely feared, rarely felt
The second-biggest fear, cognitive dependency, turns out to be mostly theoretical, according to Anthropic. Of the 56 percent who worry about it, only about one in five say they'd actually struggle if AI disappeared tomorrow. Ironically, among the 44 percent who aren't worried, roughly a third would genuinely feel the disruption.
Workers in arts and design and in education are the most concerned about cognitive dependency, according to Anthropic. Educators report observing cognitive atrophy in their students 2.5 to 3 times more often than the average respondent, according to an earlier study.

Most Americans don't want AI doing their job
About 75 percent of respondents rate AI at least as good as humans at research. Yet most still don't want it anywhere near their own work.
On most tasks, a majority of Americans did not want AI involved in their jobs, and even on the tasks they rated AI most capable—such as research and data analysis—nearly half of respondents said they want no AI involvement in their own work.
Anthropic
Acceptance does track with perceived ability, though. The better people think AI is at something, the more willing they are to let it help. And daily AI users at work are much less worried about job loss (54 percent) than people who don't use AI at all (70 percent). Hands-on experience seems to take the edge off the fear, likely because it shows both where AI can help and where it falls short, Anthropic says.
Anthropic recently published a qualitative study of 81,000 Claude users conducted with its in-house tool "Anthropic Interviewer." Job loss and cognitive dependency topped the list of concerns there, too.
— Originally published at the-decoder.com
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from The Decoder
See more →
An AI model programmed nonstop for 19 days on a single MirrorCode task that cost $2,600 to run
Epoch AI's MirrorCode benchmark reveals Claude Opus 4.7 as the leader with a 56% solve rate, reconstructing a 16,000-line toolkit in 14 hours. Despite this, all models tested struggle with the most complex tasks, highlighting limitations in current AI capabilities. The single task consumed $2,600 over 19 days, raising questions about cost-effectiveness in AI development.

