Americans Fear AI in the Workplace
A majority of US adults are worried about how AI may be used in the workplace. Anxieties are highest among older workers and those with lower incomes.
Over the last few years, we have written several NewsWires on Americans’ growing worries about the rise of AI. (See “The Limits of Techno-Optimism” and “Americans Fear AI.”) Now a new Pew survey explores workers’ attitudes toward AI in the office. The results largely speak for themselves. So let’s dive in.
Overall, workers have more negative feelings toward AI than positive ones. 55% of employed adults are “worried” about how AI may be used in the workplace, and 33% feel “overwhelmed.” Only 36% of workers feel “hopeful” and 29% “excited”—two categories, we may assume, that largely overlap.
Only 6% of workers say workplace AI use will lead to more job opportunities for them in the long run; a third (32%) say it will lead to fewer opportunities.
Feelings differ significantly by socioeconomic status. While only 30% of lower-income workers feel hopeful about AI, 45% of upper-income workers say the same. That’s a gap of 15 percentage points. Similarly, 37% of lower-income workers fear AI will lead to fewer job opportunities, while only 26% of upper-income workers say the same. That’s a gap of 11 percentage points.
The survey also found that AI use is still relatively rare in the workplace. Only 16% of workers report using AI for at least some of their tasks. In contrast, 81% of workers say little to none of their work uses AI.
Moreover, only 16% of workers say they use generative AI chatbots like ChatGPT or Gemini. The use of these tools is more common among young workers: 22% of workers ages 18-29 use chatbots at least a few times a month, compared to 11% of workers ages 50-64. The most common uses for these tools are drafting written work, editing writing, and general research.
All in all, this survey highlights the palpable fear of AI’s potential to disrupt the workforce, even though it has yet to affect many people’s actual jobs. IMO, hourly and service workers have the least to fear, while middle-skill white-collar workers face the greatest risk.
Unlike previous automation waves that primarily affected manual blue-collar jobs, generative AI is disrupting tasks that are “cognitive” but fundamentally uncreative. Many routine office jobs, which don’t require original thinking, can now be executed by bots that master complex pattern recognition. As these roles disappear, so does the need for middle management. Brookings estimates that the sectors with the greatest likelihood of AI-based automation are legal, sales, business/finance, and administrative support.
In contrast, generative AI will have a minimal effect on most service or manual labor jobs. Without a miraculous breakthrough in robotics, AI isn’t going to replace a construction worker, a firefighter, a plumber, or even a good waiter. According to the Moravec Paradox, the earlier a human task was mastered in the history of evolution, the harder it is for AI to simulate. (See “The Receding Mirage of Driverless Cars.”) Thus, the job of the busy multi-tasking farm hand will be safe from AI long after the job of the pathology-screening radiologist has been relegated to software.