A new survey of 81,000 Claude users reveals a troubling contradiction at the heart of AI's workplace transformation: people are becoming significantly more productive with AI assistance, yet those whose jobs are most exposed to the technology are also the most anxious about losing their positions. The research, conducted by Anthropic, shows that workers in occupations where Claude handles a larger share of tasks express roughly three times as much concern about AI-driven job displacement compared to those in less exposed roles.

The findings paint a nuanced picture of how artificial intelligence is reshaping work. While the average respondent reported meaningful productivity gains, with a mean rating of 5.1 on a seven-point scale (corresponding to "substantially more productive"), this optimism masks deeper anxieties among certain worker populations. One software engineer captured the sentiment bluntly: "Well like anyone who has a white collar job these days I'm 100% concerned, pretty much 24/7 concerned about losing my job eventually to A.I."

Who Worries Most About AI Replacing Their Job?

The survey data reveals a clear pattern: exposure to AI correlates directly with job displacement anxiety. For every 10-percentage-point increase in observed exposure (a measure of how much of a job's tasks Claude is used for), perceived job threat increased by 1.3 percentage points. Workers in highly exposed occupations, such as software engineers and other technical roles, expressed significantly more concern than those in less exposed fields like elementary school teaching.

Career stage emerged as another critical factor. Early-career workers, including recent graduates and junior professionals, were substantially more likely to express concern about job displacement than senior workers. This finding aligns with previous research suggesting a slowdown in hiring for recent graduates in the United States, raising questions about whether AI is reshaping entry-level job markets.

About one-fifth of all survey respondents voiced explicit concern about economic displacement. Some worried abstractly about AI replacing junior positions, while others lamented that specific aspects of their current jobs were already being automated. One market researcher noted: "In terms of improving my capability, it's no doubt. But in the future AI may replace my work".

Where Are the Biggest Productivity Gains Happening?

Despite widespread anxiety, the productivity picture is genuinely impressive across multiple income levels. The highest-paid workers, particularly software developers and managers, reported the largest productivity gains from Claude. However, some of the lowest-paid workers also described substantial improvements. A customer service representative used AI to save significant time creating responses, while a delivery driver leveraged Claude to start an e-commerce business and a landscaper built a music application using the tool.

The productivity gains varied by occupational group. Management occupations, primarily entrepreneurs using Claude to build businesses, reported the highest improvements. Computer and math occupations, including software developers, ranked second. Workers in scientific and legal professions reported the mildest productivity improvements, with some lawyers expressing frustration about AI's inconsistency when following precise instructions.

How Workers Are Using Claude to Expand Their Capabilities

• Scope Expansion: The most common source of productivity gains came from doing new tasks rather than simply completing existing work faster. Workers reported taking on responsibilities they previously couldn't handle, effectively expanding their job scope.
• Time Savings for Complex Work: High-income professionals reported dramatic time reductions on technical tasks. One developer noted that work previously taking months was completed in 4-5 days, while others described cutting four-hour tasks down to 30 minutes.
• Side Projects and Entrepreneurship: Lower-wage workers particularly leveraged Claude for technical side projects, with examples including e-commerce businesses, music applications, and other ventures that could generate additional income or career opportunities.

The survey also captured who benefits most from these productivity gains. Among respondents who indicated where benefits accrued, most cited personal advantages through faster task completion and freed-up time. However, 10% of those who named a recipient reported that employers or clients were asking for and getting more work, suggesting that some productivity gains are being captured by organizations rather than workers themselves.

The research comes as the enterprise AI market intensifies. Google Cloud recently unveiled its Gemini Enterprise Agent Platform, which includes access to Anthropic's Claude Opus 4.7, Claude Sonnet, and Claude Haiku models alongside Google's own Gemini models. This expansion of Claude's availability in enterprise platforms suggests that workplace exposure to the technology will only increase in coming months.

The paradox Anthropic's research reveals is perhaps the central challenge of AI's economic transition: the same capabilities that make workers more productive also create legitimate concerns about displacement. Workers aren't irrational to worry; they're responding to real signals about how their labor is being augmented and, in some cases, potentially replaced. The survey suggests that policymakers, business leaders, and technologists need to grapple seriously with how productivity gains are distributed and how workers in high-exposure roles can be supported through this transition.