What is the story about?
The promise of artificial intelligence has always come with a quiet, persistent question: what happens to the humans?
A new study by Anthropic, based on responses from 81,000 users of its Claude AI system, suggests that this question is no longer theoretical. It is already shaping how people think about their careers, their job security, and their place in an increasingly automated world.
The findings paint a mixed picture. On one hand, AI is making workers faster, sharper, and more productive. On the other, it is also making them uneasy, raising the spectre of a future where efficiency comes at the cost of employment.
Roughly one in five respondents, or about 16,000 people, expressed concern that AI could replace their jobs. For some, this fear remains abstract. A software developer flagged the possibility that AI might gradually eliminate junior roles, hollowing out entry-level opportunities.
For others, the anxiety is more immediate. A market researcher admitted that while AI has improved their capabilities, it may eventually render their role obsolete. In some cases, workers even reported that AI is making their jobs more demanding. One developer noted that the arrival of AI tools has led to tougher assignments, as managers expect more complex problems to be solved faster.
The study reveals a clear pattern: the more exposed a job is to AI, the greater the concern. Occupations where AI tools like Claude are frequently used, particularly in coding and technical work, show higher levels of perceived threat. By contrast, roles such as teaching, where AI adoption is lower, tend to report less anxiety.
Anthropic quantified this relationship. For every 10 percentage-point increase in AI exposure within a job, perceived job risk rises by 1.3 percentage points. Workers in the most exposed roles were three times more likely to fear displacement than those in the least exposed ones.
Career stage also plays a crucial role. Early-career professionals are significantly more worried than their senior counterparts. This aligns with broader trends suggesting a slowdown in hiring for recent graduates, particularly in the United States, where companies may be relying more on AI tools instead of expanding junior teams.
Despite these concerns, the study also highlights why AI adoption continues to surge: it works.
Using AI-assisted analysis, Anthropic rated users’ self-reported productivity on a scale of one to seven. The average score came in at 5.1, indicating that most users feel “substantially more productive” with AI. Some reported dramatic improvements, with tasks that once took months now being completed in a matter of days.
The gains, however, are not evenly distributed. Higher-paid professionals, such as software developers, reported the largest productivity boosts. This trend holds even when excluding technical roles, suggesting that jobs requiring higher levels of education benefit more from AI-driven efficiency.
At the same time, the technology is not limited to high earners. Some lower-income workers also reported meaningful improvements. A customer service representative described using AI to quickly generate responses, saving valuable time. Others are leveraging AI beyond their primary jobs. A delivery driver is using it to build an e-commerce venture, while a landscaper is experimenting with creating a music application.
Still, not everyone is convinced. A small percentage of users reported no change or even negative impacts, often citing the need for multiple attempts to achieve the desired results.
The broader takeaway is both encouraging and unsettling. AI is amplifying human capability at an unprecedented pace, but it is also reshaping expectations, workloads, and career trajectories.
A new study by Anthropic, based on responses from 81,000 users of its Claude AI system, suggests that this question is no longer theoretical. It is already shaping how people think about their careers, their job security, and their place in an increasingly automated world.
The findings paint a mixed picture. On one hand, AI is making workers faster, sharper, and more productive. On the other, it is also making them uneasy, raising the spectre of a future where efficiency comes at the cost of employment.
One-fifth of users are concerned
Roughly one in five respondents, or about 16,000 people, expressed concern that AI could replace their jobs. For some, this fear remains abstract. A software developer flagged the possibility that AI might gradually eliminate junior roles, hollowing out entry-level opportunities.
For others, the anxiety is more immediate. A market researcher admitted that while AI has improved their capabilities, it may eventually render their role obsolete. In some cases, workers even reported that AI is making their jobs more demanding. One developer noted that the arrival of AI tools has led to tougher assignments, as managers expect more complex problems to be solved faster.
The study reveals a clear pattern: the more exposed a job is to AI, the greater the concern. Occupations where AI tools like Claude are frequently used, particularly in coding and technical work, show higher levels of perceived threat. By contrast, roles such as teaching, where AI adoption is lower, tend to report less anxiety.
Anthropic quantified this relationship. For every 10 percentage-point increase in AI exposure within a job, perceived job risk rises by 1.3 percentage points. Workers in the most exposed roles were three times more likely to fear displacement than those in the least exposed ones.
Career stage also plays a crucial role. Early-career professionals are significantly more worried than their senior counterparts. This aligns with broader trends suggesting a slowdown in hiring for recent graduates, particularly in the United States, where companies may be relying more on AI tools instead of expanding junior teams.
Who benefits from AI?
Despite these concerns, the study also highlights why AI adoption continues to surge: it works.
Using AI-assisted analysis, Anthropic rated users’ self-reported productivity on a scale of one to seven. The average score came in at 5.1, indicating that most users feel “substantially more productive” with AI. Some reported dramatic improvements, with tasks that once took months now being completed in a matter of days.
The gains, however, are not evenly distributed. Higher-paid professionals, such as software developers, reported the largest productivity boosts. This trend holds even when excluding technical roles, suggesting that jobs requiring higher levels of education benefit more from AI-driven efficiency.
At the same time, the technology is not limited to high earners. Some lower-income workers also reported meaningful improvements. A customer service representative described using AI to quickly generate responses, saving valuable time. Others are leveraging AI beyond their primary jobs. A delivery driver is using it to build an e-commerce venture, while a landscaper is experimenting with creating a music application.
Still, not everyone is convinced. A small percentage of users reported no change or even negative impacts, often citing the need for multiple attempts to achieve the desired results.
The broader takeaway is both encouraging and unsettling. AI is amplifying human capability at an unprecedented pace, but it is also reshaping expectations, workloads, and career trajectories.
