If a recent survey is to be believed, managers’ use of artificial intelligence (AI) tools in making personnel decisions may be more extensive than their employers realize. A survey of 1,342 full-time manager-level employees commissioned by ResumeBuilder.com revealed some eye-popping results. Roughly 60% of respondents said they relied on AI to make decisions about direct reports, including regarding such consequential decisions as raises, promotions, layoffs and terminations. And while most reported using AI simply as a tool to assist with these decisions, 20% admitted to letting AI make decisions without human input either “all of the time” or “often.”
Many of the survey respondents appeared to be using AI tools for these purposes with little to no established limits or guidance. 24% of those using AI said they received “no training at all” on how to ethically use AI in managing people, and among those who reported at least some training, a majority described that training as “informal.” Additionally, most respondents reported primarily using general purpose chatbots for these purposes, rather than using tools specifically designed for managing employees.
While some employers may benefit (perhaps substantially) from integrating AI tools into people management, employers should be concerned about the type of use indicated by the survey, which may expose organizations to unnecessary risk. It appears that employers themselves may be unaware of the risks, as a majority of respondents indicated their company encourages them to use AI in managing their reports.
If not deployed with care, AI tools may create unintended disparate impacts that implicate equal employment opportunity laws (such as Title VII, the Americans with Disabilities Act, the Age Discrimination in Employment Act, and similar state laws), subject the employer to onerous bias audit requirements, or potentially violate other employment laws (such as state laws regarding the use of lie detector tests). For a detailed discussion of these risks, we recommend you check out our four-part podcast series, AI at Work.
Employers would be well-advised to limit the use of AI for people management only to approved uses that are clearly defined. Potential use cases should be carefully vetted (including with legal counsel) and employees should be trained regarding appropriate uses of AI tools to mitigate these risks. Employers should also establish clear policies and procedures governing the use of AI for HR functions.