Workforce and AI Literacy
Module 5 · People and Change

Workforce and AI Literacy

Plan how AI changes jobs and skills, build role-based AI literacy that meets the EU AI Act's literacy duty, and communicate honestly about AI and employment.

AI Changes Tasks First, Jobs Second

AI rarely replaces a whole job at once. It changes tasks within jobs: drafting goes faster, research takes minutes instead of hours, routine cases are handled automatically, and people spend more time on judgment, exceptions, and relationships.

That makes workforce planning a task-level exercise. For each major role family, ask:

  • Which tasks can AI automate, with little human involvement?
  • Which can it augment, making people faster or better?
  • Which stay human, because they depend on judgment, accountability, trust, or physical presence?
  • What new tasks appear: reviewing AI output, handling exceptions, managing agents, maintaining knowledge bases?
Illustrative: one customer-service role, by share of working timeAutomate 23%Routine statusquestionsAugment 40%Drafting replies,finding answersStay human 25%Complaints, empathy,judgment callsNew 12%Reviewing AI,fixing contentFreed capacity goes to:Growmore work, same teamImprovebetter quality and serviceReduceattrition or redeploymentDecide the mix deliberately, and say so.

The answers show where capacity is freed, where skills need to change, and where roles may need redesign, which is more useful than a headline estimate of “jobs affected.”


Decide What Freed Capacity Is For

When AI frees capacity, leaders have three choices:

  1. Grow: serve more customers, enter new markets, or take on work that was previously unaffordable, with the same team.
  2. Improve: use time for higher-quality work: more thorough reviews, better customer conversations, faster response.
  3. Reduce: lower headcount through attrition, redeployment, or reductions.

Most organizations will do some of each. What matters is deciding deliberately and saying so. Undeclared intent breeds suspicion, and suspicion kills adoption.

Some companies that announced large AI-driven reductions later reported quality problems and rehired people for customer-facing work. Plan for the human review, exception handling, and relationship work that AI doesn’t remove.

⚠️ Common Pitfall: Silence About Jobs

Common mistake: launching AI programs while avoiding any statement about what it means for employment.

People fill the silence with the worst assumption. Employees who believe AI is coming for their jobs won’t share the process knowledge that makes AI work, and won’t report its errors. Be as clear as you honestly can about intent, commitments (such as redeployment or retraining), and what you don’t yet know.


Role-Based AI Literacy

AI literacy isn’t one course for everyone. Different groups need different depth:

  1. 1 Everyone

    Safe use, data rules, limits

  2. 2 Daily users

    Prompting, checking, workflows

  3. 3 Managers

    Redesigning work, measuring

  4. 4 Builders

    Evaluation, security, standards

  5. 5 Executives and board

    Strategy, risk, oversight

  • Everyone: what AI is and isn’t, approved tools, what data never goes into AI, how to spot errors, and how to report problems.
  • Daily users: effective prompting, verifying output, and building AI into their workflow.
  • Managers: identifying use cases, redesigning team work, setting review standards, and measuring value.
  • Builders: evaluation, security (including prompt injection), documentation, and the organization’s AI standards.
  • Executives and board: strategy, risk appetite, regulatory exposure, and the right questions to ask.

The EU AI Act’s AI literacy duty (Article 4) requires providers and deployers to take measures to ensure sufficient AI literacy among staff and others who operate or use AI systems on their behalf, taking into account their roles and the context. A role-based program with records of who completed what is the practical way to show you’ve met it.


Building Skills That Stick

  • Learn on real work. Short sessions using people’s own tasks beat generic courses.
  • Champions networks. Enthusiastic early adopters in each team who share tips and help colleagues.
  • Communities of practice where people share prompts, workflows, and failures.
  • Manager modeling. Adoption follows managers who visibly use AI well and talk about its limits.
  • Make time. People can’t learn new ways of working with zero slack. Budget time, not just licenses.
🧪 Try It: Map One Role Family

Choose a role family in your organization (customer service, finance analysts, recruiters):

  1. List its eight to ten main tasks.
  2. Mark each task automate, augment, or stay human, and note any new tasks.
  3. Decide whether freed capacity should go to grow, improve, or reduce, and draft one honest sentence explaining that to the team.
  4. Outline the literacy each layer of the program would need for this group.
🧠 Knowledge Check

1. Why should workforce planning for AI start at the task level?

2. What's the main risk of saying nothing about AI's impact on jobs?

3. What's the practical way to show compliance with the EU AI Act's AI literacy duty?

Up Next

The next lesson covers leading adoption across the organization, including shadow AI and incentives.