AI is writing your reviews. Trust is cracking, and HR needs a response
AI-assisted writing has moved from a side experiment to an everyday habit at work. A new ZeroBounce survey of 1,000 US professionals shows employees and managers are leaning on AI for routine and sensitive communication. That includes performance reviews and layoffs. If this trend continues, traditional HR functions will shrink - and trust will go with them.
What the numbers say
- 24% of employees use AI daily to draft or edit emails, with tech workers leading the way.
- 35% have used AI to help write sensitive messages - some copy and paste with no edits.
- 20% say they've caught a coworker sending an identical AI-generated email.
- 26% suspect their performance review was generated by AI (higher among younger and tech workers).
- 16% of laid-off employees believe the termination email was written by AI; 20% said they cried after reading such a message.
- 41% of managers admit using AI to draft or revise performance reviews; 17% have used it in layoff emails.
Source: ZeroBounce survey
Why employees feel burned
People can spot template language. That erodes psychological safety and signals "you're a number."
Performance feedback feels less credible when the voice is generic, the examples are thin, or the tone is off. In high-stakes moments like layoffs, AI-written messages read cold - and employees remember that.
Why managers lean on AI
Writing is hard, time is tight, and AI provides structure and tone on demand. It reduces the pressure of crafting the "perfect" message.
The problem: convenience becomes dependency. Overuse turns leaders into distributors of stock language, not stewards of culture.
The HR risk
When sensitive messages sound automated, trust drops and attrition rises. Bias and privacy risks grow if managers paste private details into unapproved tools.
If AI proves more "efficient" than human-led HR processes, organizations will shrink teams and centralize workflows. That accelerates the decline of traditional HR - unless HR sets the rules and upskills the business.
Policy that actually works
- Draw the line: Prohibit AI for core content in performance reviews, promotions, disciplinary notes, and layoffs. Allow AI for grammar and structure only - with human rewriting and final approval.
- Disclosure: For sensitive topics, require a simple note: "This message was assisted by AI and reviewed by [Name]." Transparency restores credibility.
- Human sign-off: Any high-stakes message must be read aloud by the sender and approved by HR. If it sounds robotic, rewrite.
- Data safety: Ban uploading PII or confidential data into public models. Use approved tools with audit logs and retention controls.
- Consistency: Provide tone, clarity, and empathy guidelines with good and bad examples. Make it easy to do the right thing.
- Feedback loop: Add a one-click pulse after reviews and layoffs to monitor perceived empathy and clarity. Audit samples quarterly.
Manager playbook: ethical AI, real voice
- Start with real notes and evidence. Use AI to tighten structure, not to invent feedback.
- Replace generic phrasing with specific examples, outcomes, and next steps.
- Check for unintended tone (defensive, vague, overly formal). Rewrite in your voice.
- For layoffs and corrective actions, prioritize live conversation. The email documents; the human delivers.
Skills to build this quarter
- Responsible prompt writing for clarity, bias checks, and tone control.
- Evidence-based feedback: behavior, impact, and action (BIA framework).
- Data hygiene: what never goes into an AI tool.
Need structured upskilling for managers and HR teams? Explore role-based options at Complete AI Training.
Bottom line
AI is now part of workplace communication. The question is whether it becomes a mask or a multiplier.
HR should set guardrails, train for accountable use, and put empathy back at the center. Trust is the real KPI - write like it matters, because it does.
