Does AI help or harm? Why and how AI use influences workplace outcomes and employee well-being

Although the recent surge in excitement surrounding artificial intelligence (AI) might suggest it is a new development, AI has been applied in organizations for over forty years (Fanti et al., 2022; Markelius et al., 2024). Indeed, the term "artificial intelligence" itself emerged in the 1950s during a research project at Dartmouth College, where it was used to describe "machines able to simulate human intelligence" (Haenlein & Kaplan, 2019, p. 3). Most recently, in 2022, the landscape of AI use for laypeople greatly changed when OpenAI introduced ChatGPT, a generative AI tool, to the public; ChatGPT is an advanced AI language model designed to understand and generate human-like text based on user input (Kalla et al., 2023). Within two months of its launch, ChatGPT achieved the fastest user adoption in history, reaching 100 million active users (Hu, 2023). Alongside this rapid consumer uptake, investments in AI by businesses have been steadily increasing and are expected to double next year alone (Annual Global Corporate Investment in Artificial Intelligence, by Type, 2023; McWilliams, 2024). While research is showing that the adoption of generative AI in organizations is beginning to show value for the business, such as, greater stock returns when AI is used in business to business marketing (Zhan et al., 2024), meaningful cost reductions in human resources, and revenue increases in supply chain management (Singla et al., 2024), its impact on employees remains under-explored. Despite the excitement around AI and particularly generative AI among individual consumers and the growing investments by companies, research on AI in the workplace is limited. Some general advice exists on using AI effectively, such as a practical paper explaining how to create valuable prompts so that HRM assistants can use ChatGPT effectively (Aguinis et al., 2024). Another paper sheds light onto employee preferences, demonstrating that in work situations requiring high empathy, employees prefer human managers as opposed to AI ones; in fact, employees perceive AI management to be less benevolent and this leads to a negative impact on trust in AI management (Li & Bitterly, 2024). Some studies have begun exploring the impact of AI in the workplace on employees, but the general trend is that effects are not straightforwardly positive or negative. However, predictive studies showing why and under what conditions generative AI adoption is influencing individuals’ workplace outcomes and well-being are rare. One study looks at how employee wellbeing and performance can be helped or harmed as a function of AI use at work; the research finds that employees’ wellbeing suffers when they interact with AI a lot, because they feel increased loneliness, which leads to greater insomnia and alcohol use. This effect is greater for those with attachment anxiety (Tang et al., 2023). Yet, employees can also interact with AI a lot, feel a need for affiliation, and thus be more helpful to coworkers, thereby increasing their own contextual performance (Tang et al., 2023). Another study explores the use of AI for repetitive tasks at a telemarketing company; on one hand, it finds that employees can solve more difficult tasks creatively when they use AI for basic tasks, but this is usually for skilled employees (Jia et al., 2024). On the other hand, lower skilled employees experience negative reactions to the AI help (Jia et al., 2024). Another study finds that frequent use of AI leads to greater knowledge gain and improved task performance, but also to information overload, lowered performance and detachment from work (Shao et al., 2024). These relationships depend on employee levels of openness and positive affect (Shao et al., 2024). Overall, the few studies that explore the impact of AI at work show mixed impacts.