Is AI at Work Making Employees Lonelier?

Artificial intelligence (AI) is increasingly part of our lives: it’s used in marketing, design, finance, and manufacturing, and it makes many jobs easier. But recent studies have found that workers who deal with AI report insomnia, loneliness, and a greater likelihood of drinking after work.

Researchers at the University of Georgia ran four experiments in the U.S., Taiwan, Indonesia, and Malaysia that suggest working with AI can disrupt personal lives, according to the Daily Mail. For example, one study surveyed 166 engineers at a Taiwanese biomedical company who worked with AI systems over three weeks. The researchers measured the workers’ feelings of loneliness, attachment anxiety, and sense of belonging.

Colleagues also rated how helpful each worker was, while family members reported on insomnia and alcohol use after work. Engineers who spent more time working with AI were more likely to reach for a drink after hours, have trouble falling asleep, and feel lonely. At the same time, those same workers were more likely to help their colleagues — a behavior the researchers say may reflect a need for social interaction.

Other experiments with employees at a real estate management company and a tech firm in Indonesia, Malaysia, and the U.S. produced similar results.

Lead researcher Pok Man Tan previously worked at an investment bank and used AI systems on the job, so the topic is familiar to him. “The rapid development of artificial intelligence systems is causing a new industrial revolution that is transforming the workplace, providing many benefits but also some yet-to-be-explored dangers, including potentially harmful mental and physical effects on workers,” Tan said. “Humans are social beings, so isolation from colleagues while working with AI systems can have detrimental side effects on employees’ personal lives.”

So what can be done? Technological advancement isn’t going away. The researchers recommend that AI developers build in options that simulate human interaction — for example, a human voice — and that employers limit how long workers interact with AI and make sure they have opportunities to communicate with colleagues.