By Afla KC, Level Up HR Solutions
Let me paint you a picture that’s already happening inside your company – whether you know it or not.
It’s Tuesday morning. Sarah, a senior marketing manager, opens her laptop. She doesn’t write the first draft of the campaign brief. An AI tool does it in 12 seconds. She doesn’t summarize the competitor research. Another AI reads 40 PDFs and spits out a bullet-point table. She doesn’t even write her own code snippets anymore – GitHub Copilot finishes her sentences.
Sarah is not a robot. She’s a human. But she now works alongside three or four AI “teammates” that she invited in herself, because they save her four hours a day.
Meanwhile, her HR department still has a policy from 2019 that says “no unauthorized software.” Her manager has no idea how to evaluate her performance when she uses AI. And the company’s data security team is quietly panicking about proprietary information being fed into public AI models.
This is not the future. This is Tuesday.
And if you work in HR, you are already behind.
The Secret Workforce You Didn’t Hire
Let’s start with a hard truth: Your employees are using AI whether you approve it or not.
A 2024 survey by Microsoft and LinkedIn found that 75% of knowledge workers are already using generative AI at work. And here’s the kicker – more than half of them are using their own personal tools, not company-approved ones.
They’re feeding customer emails into ChatGPT to draft responses. They’re asking Midjourney to create presentation images. They’re using Otter.ai to transcribe and summarize meetings they didn’t attend.
This is the shadow AI workforce – and it’s growing faster than any contingent labor pool you’ve ever managed.
Why are they doing it? Not because they’re lazy. Because they’re overwhelmed. Because their to-do lists have doubled while their calendars have shrunk. Because AI is the first tool in years that actually makes them feel competent again.
If HR ignores this, two things happen:
- Massive data risk. Proprietary information leaking into public AI models is already a real problem. Samsung employees accidentally leaked sensitive code via ChatGPT within weeks of its launch.
- Massive equity problem. The employees who are already confident, tech-savvy, and plugged into online communities get superpowers. The ones who are burned out, less connected, or intimidated? They fall further behind.
Your job as HR is not to ban AI. Your job is to manage the transition – just like you managed the transition to remote work, to Slack, to mobile email.
The Performance Review Breaks (Again)
Remember how performance reviews
broke when everyone went remote? Same thing is happening now – but worse.
Let’s be honest: How do you evaluate someone who uses AI to do 80% of their work in half the time?
Old model: “Sarah wrote a 10-page report. Good output. High effort.” New model: “Sarah prompted an AI to write a 10-page report in 20 minutes, then spent 2 hours fact-checking, adding unique insights, and making it sound like her. Is that less valuable? More valuable?”
I’ve sat in five HR roundtables this year where no one could answer that question cleanly.
Here’s what I’m starting to believe: We need to stop measuring effort and start measuring value-added.
If an AI can do 80% of a task, the human’s job is the 20% that the AI cannot do: judgement, emotional intelligence, ethical reasoning, creativity that breaks patterns, relationships.
But most job descriptions today don’t even mention those things. They say “write reports”, not “curate AI-generated content for strategic accuracy”. They say “analyse data”, not “design prompts and validate AI outputs”.
Your performance management system is built for a world that no longer exists.
And if you don’t redesign it, you’ll end up demoralising your best people – the ones who figured out how to work smarter – while celebrating the ones who grind inefficiently.
The Skills Revolution No One Is Talking About
I want to introduce you to a new term: prompt literacy.
Prompt literacy is the ability to talk to AI in a way that gets useful, accurate, and safe results. It’s not coding. It’s not data science. It’s a new form of communication – part search engine, part negotiation, part creative writing.
Right now, prompt literacy is distributed completely unevenly across your workforce. The 25-year-old who grew up on Reddit? Probably great at it. The 52-year-old operations manager who still prints emails? Probably not.
That gap is not age-related. It’s exposure-related. And left unaddressed, it will become the next digital divide – wider and more corrosive than any we’ve seen.
So what does HR do?
You don’t need to teach everyone Python. But you do need to:
- Offer hands-on AI literacy workshops (not theory, not ethics lectures – actual prompting exercises)
- Create internal prompt libraries where teams share effective prompts for common tasks
- Normalize “AI co-pilot” as a job skill – put it in job descriptions, learning paths, and performance criteria
And here’s the radical part: Encourage people to admit when they use AI.
Right now, many employees hide it. They feel like using AI is cheating. That fear is deadly. It kills transparency, kills learning, and kills the chance to set guardrails.
When someone says “I used ChatGPT to help draft this email,” the right response is: “Great. Show me how. Then let’s talk about what you changed and why.”
Data Privacy, Bias, and the New HR Compliance Nightmare
Okay, let’s talk about the messy stuff.
Every time an employee pastes a customer list into a free AI tool, that data is likely being used to train the model. Every time they ask an AI to summarize a termination discussion, they’re potentially violating privacy laws. Every time they rely on an AI to screen a resume (yes, people are doing this too), they’re inheriting whatever bias the model was trained on.
Most companies have zero policies around this.
I’m not saying you need a 50-page AI governance document. That will gather digital dust. But you do need three things:
1. A simple “stop, think, then prompt” rule
Create a one-page guide with red/yellow/green zones:
- Red (never enter): PII, trade secrets, HR records, legal documents
- Yellow (enter only with approved enterprise tool): internal strategy, financial projections, customer data
- Green (safe): public information, brainstorming, summarizing public articles
2. An approved enterprise AI tool
Instead of banning all AI, buy one enterprise-grade tool (e.g., ChatGPT Enterprise, Microsoft Copilot, Google Gemini for Workspace). It costs money. It also gives you data controls, audit logs, and legal protection. Consider it the cost of not having a data breach.
3. A bias-checking protocol for any AI used in people decisions
If anyone – recruiter, manager, HRBP – uses AI to evaluate candidates, write performance feedback, or suggest promotions, they must also use a second tool or human review to check for bias. Make it a mandatory two-step.
This is not paranoia. This is the same care you already take with spreadsheets and email. AI just amplifies speed – both of good decisions and bad ones.
The Human Skills That Actually Matter Now
Here’s the hopeful part.
For years, we’ve been told that robots will replace us. And some tasks will absolutely be automated. But the more I watch AI evolve, the more I see human skills becoming more valuable, not less.
AI can write a decent email. It cannot build trust after a layoff. AI can summarize a meeting. It cannot read the room and sense that someone is about to cry. AI can suggest a project plan. It cannot inspire a burned-out team to care again. AI can analyze diversity data. It cannot sit with a junior employee and say “I see you, and I’ve got your back.”
The premium on empathy, judgement, ethical courage, and genuine connection is about to skyrocket.
So when you think about “managing the AI-augmented workforce,” don’t think about controlling the machines. Think about liberating the humans – from drudgery, from low-value busy work, from the soul-crushing parts of their jobs – so they have energy for the work that only people can do.
That’s the real frontier.
What HR Leaders Should Do This Quarter
Enough theory. Here’s a 90-day action plan.
Month 1: Listen and learn
- Run an anonymous survey: “What AI tools are you using for work right now?”
- Host three brown-bag lunches where people share their AI wins and fears
- Talk to your legal and security teams – what’s the actual risk level?
Month 2: Create simple guardrails
- Publish the red/yellow/green data rule (one page, not 50)
- Pilot one enterprise AI tool with one department (e.g., marketing or customer support)
- Write a one-paragraph AI use policy – short enough to fit on a sticky note
Month 3: Build capability and culture
- Run live prompting workshops (not videos, not PDFs – actual hands-on)
- Update one job description to include “AI literacy as a plus”
- Start a weekly “AI office hour” where people can ask dumb questions without shame
You don’t have to solve everything. You just have to start. Because your employees already have.
A Letter to the HR Leader Who Feels Overwhelmed
I know what you’re thinking.
“I can’t even get managers to do performance reviews on time. Now I have to manage AI?”
“We have no budget for enterprise tools. Our ATS is from 2016.”
“I’m not a technologist. I barely understand how ChatGPT works.”
I hear you. And I’m not saying this is easy.
But here’s what I also know: You have managed remote work. You have managed hybrid chaos. You have managed return-to-office wars. You have managed a pandemic, a great resignation, a quiet quitting epidemic, and now a raging debate about culture fit and DEI.
You are the most resilient, adaptable, creative function in any organization. You have learned more in the last five years than most professions learn in a decade.
Managing AI is not another crisis. It is the same muscle you’ve been building all along: helping humans work better with new tools.
You’ve got this. Really.
Let’s End With a Question
I want you to close your eyes and imagine your organization two years from now.
Every knowledge worker has an AI co-pilot. Routine tasks are 80% automated. Meetings are shorter. Email volume is down. People have time to think, to connect, to innovate.
What does your HR department need to do today to make that future real – instead of a chaotic mess of shadow AI, burned-out employees, and security breaches?
Now open your eyes.
Pick one thing from this post. Do it this week. Not next quarter. This week.
And then come back and tell me how it went.