According to Anthropic's survey of 51,993 Americans conducted in late 2025, job loss ranked as the top fear at 64%, leading in every state. Concern levels varied from 71% in Iowa to 57% in Mississippi, with Democrats at 67% and Republicans at 62%. The survey found that Americans with postgraduate degrees are nearly 10 percentage points more worried about job loss than those with high school education or less. Notably, people who use AI at work daily showed 54% worry compared to 70% for those who don't use AI at all. Recent economic analysis suggests these fears may be justified, as AI could automate every task that a human can do, potentially forcing workers to their second-best jobs and leading to wage declines.
On the positive side, 48% of respondents placed curing diseases like cancer or Alzheimer's in their top three uses for AI. Helping people with disabilities followed at 36%. The survey revealed that 71% of Americans want government involvement in AI, including 79% of Democrats and 68% of Republicans. When asked about keeping AI development aligned with humanity's interests, 47% backed holding companies legally liable for harm, while 44% wanted safety prioritized over growth. Recent medical innovations like AlphaFold and protein folding demonstrate AI's potential for delivering on these healthcare promises.
The survey findings align with concrete employment data showing AI's impact on jobs. According to BeInCrypto reports, AI drove 38,579 US job cuts in May, representing about 40% of the month's total. For 2026, employers have tied 87,714 cuts to AI, with this total already exceeding the 54,836 cuts attributed to the technology across all of 2025. The pressure has reached Washington, with Senators Elizabeth Warren and Bernie Sanders urging Congress to protect workers. However, Amazon founder Jeff Bezos rejects the job-loss narrative, predicting AI will create labor scarcity instead. Recent economic modeling suggests this optimistic scenario may not materialize for all workers, as AI could fundamentally reorient the economy and create new challenges for employment.
The survey found that only 15% of Americans trust AI companies to steer the technology, making it the lowest figure for any institution tested. This trust level was below the federal government (20%), state and local government (19%), and international bodies (20%), and far below independent experts (43%). The findings expose a gap between what the public wants from AI and what it dreads, as real layoff data and political pressure over automation build across the United States. Recent research suggests this lack of trust may be well-founded, as AI models can be easily jailbroken and could potentially be used for malicious purposes like designing deadly viruses or hacking critical infrastructure.
