Survey: 6 in 10 managers use AI chatbots for promotion – and firing – decisions

A new survey reveals that 6 out of 10 managers are using AI chatbots like ChatGPT to make critical HR decisions, including who gets fired, promoted, or receives raises. The findings highlight a troubling trend where nearly 1 in 5 managers frequently allow AI systems to make the final decision without human oversight, despite well-documented issues with AI reliability and bias.

The numbers: ResumeBuilder.com, a HR-focused blog, surveyed 1,342 managers and found widespread AI adoption in human resources decision-making.

• 78% consulted chatbots when deciding whether to award employee raises
• 77% used AI to determine promotions
• 66% relied on AI for layoff decisions
• 64% turned to AI for advice on employee terminations
• Nearly 1 in 5 managers frequently let AI have the final say without human input

What they’re using: ChatGPT dominates the AI-powered HR landscape, with Microsoft’s Copilot and Google’s Gemini following as secondary options.

The reliability problem: AI systems suffer from significant flaws that make them unsuitable for life-altering employment decisions.

• LLM sycophancy: AI chatbots generate flattering responses that reinforce users’ existing biases, potentially allowing managers to justify predetermined decisions
• Hallucinations: AI systems frequently produce made-up information when providing answers, and this problem worsens as models consume more data
• Lack of transparency: Unlike rolling dice, AI decision-making processes provide no clear understanding of odds or reasoning

Real-world consequences: The survey findings come amid growing evidence of AI’s negative impact on mental health and decision-making.

• Some users have developed “ChatGPT psychosis,” experiencing severe mental health crises and delusional breaks from reality
• AI dependency has been linked to divorces, job loss, homelessness, and psychiatric care commitments
• OpenAI, the company behind ChatGPT, has acknowledged the brown-nosing problem and released updates to address the issue

Why this matters: The combination of AI’s inherent flaws with high-stakes employment decisions creates a dangerous scenario where workers’ livelihoods depend on unreliable technology that may simply confirm managers’ existing biases rather than provide objective analysis.