A new survey from ResumeBuilder.com shows a troubling change happening in workplace management: the increasing reliance on artificial intelligence, particularly large language models (LLMs), to guide major human resources decisions.
The poll of managers across various industries revealed that a majority are using AI tools like ChatGPT to weigh in on matters such as layoffs, promotions, and salary adjustments.
There has been a sharp rise in AI’s role within HR operations. According to the data, 78% of respondents said they had consulted a chatbot when determining raises, while 77% did so for promotion decisions.
Even more strikingly, 66% acknowledged using an LLM to help decide which employees to let go. Nearly 20% of those surveyed admitted they often defer final decisions entirely to the AI, bypassing human judgment.
ChatGPT is the most commonly used AI assistant among the tools mentioned, followed by Microsoft’s Copilot and Google’s Gemini. The enthusiasm for these tools highlights their perceived utility, but it also raises critical concerns about bias, reliability, and the diminishing role of human oversight in deeply personal workplace outcomes.
Experts have previously flagged issues with LLMs reinforcing user biases through overly agreeable or flattering responses, which is a problem that has affected ChatGPT in particular.
OpenAI has implemented updates to address these tendencies, but skepticism remains about the models’ ability to make sound, objective decisions in sensitive contexts.
Beyond the office, overreliance on AI has been linked to serious mental health issues among some users. The term “ChatGPT psychosis” has emerged to describe a detachment from reality reportedly experienced by individuals who excessively depend on the chatbot for advice and decision-making.
In extreme cases, this dependence has been associated with job losses, relationship breakdowns, and psychiatric episodes requiring professional intervention.
As AI tools continue to gain traction in business environments, the findings serve as a cautionary tale about the potential risks of handing over high-stakes decisions to machines. While these technologies offer convenience and efficiency, their integration into matters of employment and personal well-being demands careful regulation and ethical scrutiny.
