Listen to this article with a free account

Americans are increasingly encountering and using artificial intelligence technologies like Anthropic’s Claude, OpenAI’s ChatGPT and Google’s Gemini. Most surveys now suggest more than two-thirds of the population is interacting with these technologies.

However, a defining question is how far along we are in a transition from experimentation with AI to integration into daily personal and professional life. How many Americans are using it regularly? For what purpose? And what value are they extracting?

One compelling way to track this adoption of AI is via Gallup polling, which found that 12% of Americans now report using AI daily at work. While the number seems modest, it represents a threefold increase in just over a year, from 4% in mid-2024.

Employees in the knowledge-work sector are leading this change. In the technology industry, 31% of workers reported daily usage. That’s compared to 19% of finance industry workers and 16% of employees in the professional service industry.

Gallup also identified the ability to work remotely as an indicator of AI adoption. Daily usage among remote-capable jobs sits at 19%, more than double the 7% daily adoption rate in roles requiring an on-site or physical presence.

For historical perspective, Pew Research Center tracked internet adoption in the early aughts. In 2000, 14% went online daily for job-related research. In 2001, the number jumped up to 19%.

While AI usage by white-collar workers is more pronounced than among blue-collar workers, pessimism regarding AI’s impact on the labor market cuts across the typical dividing line for social and cultural issues.

An Economist/YouGov poll found 63% of American adults thought that advances in AI would lead to an overall decrease in jobs. There was little difference in opinion based on education: 67% of those with college degrees and 61% of those without degrees shared this concern.

This pessimism does appear to be greater than concerns about computers in the workplace at the turn of the century. In 1999, a NPR/Kaiser/Harvard Technology survey found that 32% thought computers would lead to a decrease in jobs, while 43% thought they would lead to an increase and 23% thought computers wouldn’t make a difference.

That means today’s anxiety about AI is nearly twice as high as the computer anxiety of the late ’90s, as overall attitudes toward AI seem to resemble the mix of skepticism and curiosity seen during the early rise of the internet.

As use of AI spreads, another question is how people are actually using it. OpenAI released a study last year that estimated 30% of ChatGPT usage was work-related and 70% was personal.

The study categorized ChatGPT usage into three main behaviors. There was “asking” (49%), or prompting the AI for advice or information on a specific topic. “Doing” (40%) involved practical tasks such as drafting text or writing code. And “expressing” — chatting or playing — accounted for 11% of usage.

Writing and seeking guidance are among the leading practical uses of ChatGPT, according to OpenAI. When specifically looking at the workplace, the study found technical help and writing become even more dominant uses.

Still, a majority of respondents (53%) in a recent Fox News poll said AI had not yet made a significant difference in their lives, while 26% said it had personally helped them and 20% believed it had caused them harm.

With daily AI use in the workplace tripling in a year, it’s possible those numbers could shift even more rapidly in the months ahead, as the technology moves from the periphery and toward the center of American life.