A new study from Anthropic indicates that the impact of artificial intelligence (AI) on employment may be more complex than often suggested. The latest edition of the Anthropic Economic Index, released last week, analyzed approximately two million anonymized conversations from its AI product, Claude, over the past year, examining how AI is being utilized in the workplace and the implications for job roles.
The study found that rather than resulting in widespread job destruction, AI is leading to what the report describes as job fragmentation. According to the findings, 49% of U.S. jobs now involve tasks where AI can be employed for at least a quarter of the work involved, a significant increase from 36% in early 2025. This trend suggests that AI is increasingly taking over specific components of various jobs.
“Combining across reports, 49% of jobs have seen AI usage for at least a quarter of their tasks,” the report states. The study classified interactions with Claude by purpose—work, education, and personal use—and explored how AI is being integrated into work tasks. Some users leveraged Claude for end-to-end tasks, such as translating text and summarizing documents, while others engaged more iteratively, utilizing AI to draft, refine, or explore concepts alongside human judgment.
On Claude’s consumer-facing platform, 52% of work-related interactions involved a form of augmentation, where AI supported rather than replaced human labor. This indicates a slight shift over time, with augmentation being more prominent earlier in the year but showing signs of decreasing as automation gains traction, especially in enterprise environments.
“Data from November 2025 points to a broad-based shift back toward augmented use on Claude.ai: The share of conversations classified as augmented jumped 5pp to 52% and the share deemed automated fell 4pp to 45%,” noted the report. Changes in the product, including file creation capabilities and persistent memory, appear to have influenced these evolving usage patterns toward more collaborative interactions.
Peter McCrory, Anthropic’s head of economics, remarked that “AI is spreading across the U.S. faster than any major technology in the past century.” He noted that while adoption has been concentrated in a few states, there are signs that less engaged areas are rapidly catching up.
The impact of AI on various occupations is notably uneven. In fields like data entry and certain IT support roles, AI appears to be absorbing a greater share of core responsibilities. Conversely, in professions requiring complex judgment or interpersonal skills, AI is often used to handle routine tasks, allowing workers to focus on higher-value activities. For instance, radiologists can devote more time to patient-centered decisions rather than documentation, and therapists can reduce administrative burdens to enhance client engagement.
The study also highlights that AI tends to provide the most significant productivity improvements in complex tasks, paradoxically the same areas where it is most likely to falter without human oversight. McCrory noted that “the most complex tasks that people use Claude for are the ones where Claude tends to struggle most,” emphasizing that human guidance remains critical.
This dynamic poses a challenge to prevailing narratives about AI’s role in labor markets. While human involvement can slow productivity gains, it can also enhance them, allowing workers to amplify their output rather than face outright replacement.
The findings come amid cautionary statements from Anthropic leadership, including CEO Dario Amodei, who warned in 2025 that AI could eliminate up to half of entry-level white-collar positions and push unemployment rates into double digits within a few years. At the recent World Economic Forum in Davos, Amodei asserted that AI is “6 to 12 months” away from taking over software engineering roles.
Although the new data does not directly contradict these concerns, it portrays a more nuanced picture of AI’s economic effects, at least for the time being. Anthropic acknowledges its vested interest in portraying AI as transformative, and the rapid advancements in technology mean that current findings might quickly become outdated as automation extends further.
Ultimately, the Anthropic Economic Index provides a snapshot of a labor market in flux, with AI becoming increasingly integrated into daily work processes. Jobs are evolving into smaller, task-oriented roles, which may permanently alter the landscape of employment. While the anticipated job apocalypse may not be imminent, the nature of work is undoubtedly shifting in ways that could be difficult to reverse or accurately predict.
See also
Vizient Report: Aging Population and AI Investment Shift Hospital Economics by 2035
AI Can Generate Prose, but Memoir’s Truth Remains Beyond Its Reach, Experts Warn
Razer Unveils Project AVA: 3D AI Companion Raises Privacy and Dependency Concerns
Core Scientific Transforms to AI Infrastructure; Analysts Upgrade to Buy Amid Risks
Germany”s National Team Prepares for World Cup Qualifiers with Disco Atmosphere
