According to Inc, new data from ZeroBounce reveals that 56% of U.S. employees and managers use AI for workplace writing anywhere from a few times per week to “always,” with just 20% never using it. The survey of 1,000 people found that 35% have used bots to compose sensitive messages, including 41% of executives who admitted using AI to draft or revise performance reviews. Even more concerning, 24% of managers used apps to compose performance warnings, and 17% used AI to inform employees they were being terminated. On the receiving end, 26% of employees believe their performance reviews were AI-written, while 16% suspect their layoff notices were bot-generated. The research shows 10% of users have become so dependent on AI that they no longer feel capable of writing sensitive reports or even regular email without assistance.
The AI Empathy Gap
Here’s the thing about using AI for difficult conversations: it’s painfully obvious. Over a fifth of respondents reported catching coworkers using the exact same AI email they’d seen before. The tells include excessive formality, unnatural word choices, and that overly polished, repetitive phrasing that bots are infamous for. Basically, when you’re delivering bad news, people can smell the algorithm from a mile away.
And let’s be honest – who wants to get fired by ChatGPT? There’s something deeply unsettling about being let go by something that doesn’t actually understand what “letting you go” means in human terms. The Preply survey found that phrases like “letting you go” appeared in 45.6% of layoff announcements, while “effective immediately” and “terminating your employment” showed up in over 28% of cases. When everyone’s using the same corporate-speak, it screams “I didn’t bother to think about this personally.”
Why Managers Keep Choosing AI Anyway
So why are managers turning to bots for the most human parts of their jobs? The Preply survey offers some clues: 55% of managers who’ve fired someone never received training on how to do it. When you’re unprepared and uncomfortable, AI becomes the path of least resistance. It’s easier to let a bot handle the awkward phrasing than to stare at a blank screen trying to find the right words for terminating someone’s employment.
But here’s the problem with that thinking. When 65% of laid-off employees say their manager handled the situation poorly, maybe the solution isn’t better AI – it’s better managers. The technology is becoming a crutch for poor leadership development. Companies are investing in AI tools while apparently skipping the fundamental training on how to have difficult conversations like adults.
Where Human Judgment Still Matters
The survey participants seem to understand this balance. 40% of employees think sensitive emails about promotions and terminations should never be AI-assisted. Another 56% are okay with some AI support, but only if it doesn’t replace full human input and personalization. That’s the sweet spot – using AI to catch typos or improve clarity, but keeping the heart and judgment human.
Look, I get it. AI is incredibly useful for drafting routine emails or cleaning up messy writing. But when it comes to performance reviews, warnings, or layoffs, we’re talking about moments that can change someone’s career trajectory – or life. These aren’t tasks you want to outsource to something that’s essentially a very sophisticated autocomplete. The emotional intelligence required simply isn’t there yet, and employees can tell the difference immediately.
Where Do We Go From Here?
This isn’t about abandoning AI entirely. The technology clearly has value for automating repetitive tasks and improving efficiency. But we need to recognize its limitations, especially in areas requiring genuine human connection. Companies should establish clear guidelines about when AI assistance is appropriate and when human judgment must prevail.
Maybe the real issue here is that we’re expecting AI to solve problems that are actually about leadership development and communication skills. If managers need AI to write performance reviews, perhaps the solution isn’t better AI – it’s better management training. After all, no algorithm can replace the nuanced understanding that comes from actually working with someone day after day.
