More than two years have passed since ChatGPT was first made public, and the workplace is still adjusting to artificial intelligence.
Some hints about how AI is changing work can be found in two surveys that were either published or carried out last month, as well as one extensive field experiment.
The research indicates that some employees are dubious and cautious, even though AI has demonstrated promise in bridging experience gaps and enhancing teamwork. This is especially true when there is little or no direction on how to use AI.
This is what the researchers discovered.
AI can be a helpful team member
AI-augmented teams were much more likely to generate high-quality solutions, according to a working paper, and workers with AI performed on par with two-person human teams.
In a pre-registered, randomized controlled study, researchers from the Digital Data Design Institute, Wharton, and Harvard gave 776 Procter & Gamble personnel real-world product development tasks to complete. They included coming up with concepts for packaging and suggesting retail tactics, either with or without GPT-4 or GPT-4o’s assistance.
One of four groups—working alone or in teams, with or without access to AI—was assigned to participants at random. The study assessed emotional reaction, task completion time, and solution quality.
According to the study, teams with AI fared better overall and were more likely to provide the top 10% of answers than teams with only human assistance. Individuals utilizing AI performed on par with human teams operating without AI.
In a March blog post, Ethan Mollick, a professor at UPenn’s Wharton School and one of the paper’s authors, stated that the results indicate “AI sometimes functions more like a teammate than a tool” and that businesses should reconsider the technology.
According to him, companies should reconsider how they distribute cross-functional assignments, train employees, and organize teams.
Employees with less expertise in product development performed on par with teams of experts when given access to AI, according to the study, which suggests the technology can assist bridge functional knowledge gaps.
According to the study, using AI at work enhanced the emotional experience. Compared to those who did not use AI, those who did reported feeling less anxious and frustrated and more excited, energized, and enthusiastic.
Compared to those working in human teams, people using AI occasionally had more favorable emotional reactions.
The paper is currently at the working draft stage and has not yet undergone peer review, even though the experiment only took one day to complete. The study relied on GPT-4 models and only small teams completed tasks over the course of one day, which the researchers recognize are limitations.
Most Gen Z workers say they lack AI guidance
A survey conducted in March by Gallup and the Walton Family Foundation revealed that the majority of Gen Z employees, while being digital natives, are navigating AI at work without clear guidance.
55% of Gen Z employees reported that their companies had no official AI policy. Even among those who do have regulations at work, just roughly 10% say that they are “extremely clear.”
3,465 Americans aged 13 to 28 who lived in all 50 US states participated in a Gallup web survey, which found that Gen Z workers and students who work in settings with clear AI policies are more likely than their counterparts to consistently use AI.
Many younger workers, however, have doubts about the caliber of work produced by AI. According to nearly two-thirds of Gen Z respondents, they would be more inclined to trust work created by humans than work completed or aided by artificial intelligence.
The results show that Gen Z as a whole has mixed feelings about AI’s influence on their ability to think critically and their productivity at work or education. Of Gen Zers, 49% said AI would harm their capacity for critical thought, compared to 22% who believed it would improve it.
Seventy-two percent of respondents said AI may make it easier for them to find information, and sixty-six percent said it could speed up their work.
The debate over AI divides software engineers
Vibe coding, which uses AI to generate software through prompts, is revolutionizing the software development industry. The movement does not seem to be supported by all software engineers, though.
In a March Wired study of 730 software professionals, three out of four developers reported using ChatGPT or other AI tools, and the majority reported using AI at least once a week.
However, different people had different levels of enthusiasm. Approximately 40% of respondents described themselves as “AI pessimists,” whereas slightly more than one-third named themselves optimists. Nearly half of mid-career programmers expressed concern about the effects of AI, making them the most skeptical group.
Developers in their early careers, however, seemed more upbeat. Seventy-five percent of engineers who had less than a year of coding experience described themselves as AI optimists. Nearly one in three seasoned programmers with more than 20 years of experience also reported using AI into their operations.
However, not every adoption is public. Four out of five full-time programmers reported using AI at work without disclosing it to their employer.
This uneasiness may be a reflection of more general concerns about the job situation. Job posts for software engineers in the US have decreased by a third on Indeed since five years ago.
