In Part 2, I offered a better approach: Use AI to augment, not erase.
Now let’s make it real.
AI Use Case or Ethical Red Flag? Ask These 3 Questions.
Every leader I know is being pitched new AI use cases—automate this, optimize that, cut costs across the board. But before you approve, I suggest running the idea through this quick test:
✅ 1. What’s the real purpose—innovation or elimination?
If your primary KPI is reducing headcount, you may be automating a problem, not solving it. Ethical AI projects should be driven by improving outcomes, boosting quality, or expanding access, not by shrinking payroll.
- Good sign: The use case frees your team to focus on higher-value work.
- Red flag: The pitch starts with how many people you can let go.
✅ 2. Does it preserve or erode human dignity?
Some work shouldn’t be handed off to algorithms, especially when it involves empathy, judgment, or trust. Think of teachers, nurses, social workers, or support staff with deep institutional knowledge. I’ll include the recently displaced federal workers and contractors to this list.
- Good sign: The AI system supports human decision-making and respects context.
- Red flag: It replaces human judgment in emotionally complex or high-risk environments.
✅ 3. Could you defend this decision in the open?
If your team, customers, or the public found out how you’re using AI, would they respect it, or resist it? Ethical decisions can withstand scrutiny. If you’d rather keep the use case under wraps, it may not pass the smell test.
- Good sign: You can explain the decision proudly and clearly.
- Red flag: The use case wouldn’t survive a public hearing or all-hands meeting.
Ethical AI isn’t a checklist—it’s a mindset.
It’s not about avoiding innovation. It’s about owning the impact of your choices. Because when we move too fast, hide behind automation, or treat people as cost centers, we don’t just risk bad outcomes. We risk losing who we are as leaders.
Let’s raise the bar for AI adoption—so our teams, our companies, and our communities don’t pay the price later.
What questions do you ask when evaluating AI use cases?
