What makes this study significant is that it moves beyond predictions and surveys to show exactly how people are interacting with AI on the job.
Workers are primarily using AI as a tool to help with specific tasks throughout the workday, rather than to fully automate what they do, according to a new report from Anthropic, the company behind AI assistant Claude.
Artificial augmentation: Based on an analysis of how users interacted with Claude, the report found 57% of AI usage could be classed as "augmentation," where workers used the technology to refine ideas, brainstorm, and check for accuracy. The other 43% of interactions involve direct automation, where AI was used to complete tasks with minimal human input.
No jobs takeover (yet): Despite concerns about job displacement, the report suggests that AI is enhancing productivity rather than replacing workers. Only about 4% of occupations see AI used for 75% or more of their tasks. However, 36% of occupations use AI for at least a quarter of their tasks, indicating broad adoption across the workforce.
Popular with writers and coders: The areas of work where AI adoption was highest software development and content creation. Software engineering leads AI adoption, accounting for 37.2% of AI interactions. Common use cases include debugging, modifying code, and troubleshooting. In contrast, industries requiring physical labor, such as healthcare, transportation, and agriculture, have significantly lower AI adoption.
An expert’s take: "What makes this study significant is that it moves beyond predictions and surveys to show exactly how people are interacting with AI on the job," said Conor Grennan, Chief AI Architect at NYU Stern School of Business. In a LinkedIn post, Grennan outlined his three key takeaways from the study:
Signs of a workforce shake-up: The report suggests that while AI is not yet replacing jobs at scale, its increasing use to carry out high-value tasks indicates a looming significant impact on workforce dynamics.
"AI has already begun to diffuse into task portfolios across a substantial portion of the workforce," the report said. "While our data reveals where AI is being used today, inferring long-term consequences from these early usage trends poses significant empirical challenges."
