Ethical Leadership in the AI Era
Leadership has changed a great deal in the last twenty years. Many of us now lead teams across locations and screens, and we work with tools that can draft a report or recommend a decision in seconds. The ethical questions surrounding digital technology aren't new, but today's AI tools have made many of them more immediate and more common in everyday leadership.
Technology can produce an answer in moments, but we can't rely on a tool to independently guarantee that its own output is accurate or appropriate, and it can't take responsibility for the result. Those responsibilities remain ours. Leaning on a polished answer in place of our own judgment is easy, and so is letting messages and dashboards replace real conversations with the people we lead.
Ethical leadership in the AI era requires us to use technology thoughtfully while verifying what we're given, owning our decisions, applying discernment, and staying connected to the people we lead.
Self-Assessment:
Ethical Leadership in the AI Era
Please take a few moments to reflect on the following questions. Where can you identify opportunities for growth in your leadership?
Before I share or act on information I find online or get from AI, do I verify it?
Can I explain and stand behind every decision technology helped shape?
Before I turn to a search or AI for help, do I identify what I already know, what I need to find out, and what assumptions I may be bringing in?
Does a (seemingly) polished answer ever keep me from questioning its accuracy?
Do I review information more carefully when a decision could have serious consequences?
With team members I rarely see, do I still make time for real conversations?
Do I understand how confidential information is handled when I enter it into a digital tool?
Are the people I lead aware of when and how AI influences decisions affecting them?
This self-assessment is a starting point for understanding Ethical Leadership in the AI Era as a leader. Reflect on your responses, identify areas for growth, and use feedback from others and your ECFL Leadership Coach to guide your development.
For most of history, the challenge was finding information. Today, answers are a few keystrokes away, and the challenge lies in knowing which answers to trust. A search result, a social media post, a forwarded article, or an AI-generated summary can seem accurate at first glance, whether it is or not.
Because information can be produced and shared so quickly, it's easy to move from receiving an answer to acting on it without stopping to check its reliability. We can also give more attention to information that supports what we already believe while overlooking evidence that challenges it.
Giving people access to technology is only the first step. Using it well requires knowing how to question, verify, and apply what it produces. Leaders remain responsible for the decisions they make with this information, and the goal is better decisions made with integrity. Leaders also set expectations for how technology should be used, including what information to question, what can be shared, and which tools or channels are appropriate.
“Integrity without knowledge is weak and useless, and knowledge without integrity is dangerous and dreadful.”
When information is abundant, discernment helps us decide what deserves our attention and what we can trust.
Discernment is the ability to weigh information, understand the context around it, question the assumptions behind it, and apply human judgment to the decision. In his review of research on managerial discernment, researcher Kent D. Miller highlights three behavioral components identified in earlier work: knowledge acquisition, self-regulation, and knowledge application. (LEAD, a simple method we'll look at next, turns this broader idea into a practical check we can use before we act.)
Not every answer requires the same level of scrutiny. A routine, low-stakes task may call for only a brief pause, while a decision that affects someone's job, safety, privacy, or well-being deserves much closer review. The greater the consequences of getting it wrong, the more carefully we should apply discernment.
Have you ever accepted information because it sounded confident or matched what you already thought?
Researchers at Harvard Business School, working with MIT Solve, studied how people respond to AI recommendations in a real setting. In a field experiment, 228 evaluators screened 48 early-stage proposals submitted to a global health challenge, working in one of three ways: without AI, with an AI recommendation to pass or reject each proposal, or with the same recommendation plus a written explanation. AI recommendations improved evaluators' decisions against an independent expert benchmark compared with working alone. The written explanations made the recommendations more persuasive without making the decisions more accurate, since evaluators went along with the AI more often when it explained itself and their results were no better than those of evaluators who saw only the recommendation. The researchers describe narrative coherence as a decision shortcut, meaning a well-written explanation can substitute for judging the proposal itself.
“Discernment is not a matter of simply telling the difference between right and wrong; rather, it is telling the difference between right and almost right.”
A separate study of 666 participants in the United Kingdom found that frequent use of AI tools was associated with lower measures of critical thinking, with cognitive offloading, the habit of handing mental effort to a tool, helping to explain the link. The study reports a correlation and relied on surveys and interviews, so it can't show that AI use causes weaker thinking; younger participants also showed higher dependence on AI tools and lower scores than older participants.
Both findings point to the same habit: when an answer arrives polished and fast, our own scrutiny can relax. LEAD keeps scrutiny in place by turning discernment into four steps:
Look for what's missing: Ask what context and human experience an answer leaves out. The people closest to the situation hold the details no feed or tool can see, such as a client's history or a team's mood.
Examine the facts: Verify the key claims, and trace each one to its source. Information spreads rapidly, so it helps to understand how it was created and how it reached us, since a claim repeated widely can still be wrong.
Ask about assumptions: Identify what the answer takes for granted, and look for bias or errors, including assumptions built into how the question was asked. Consider whether an ethical concern, competing responsibility, or affected person has been left out of the reasoning.
Decide and own it: Treat the answer as one input among several, and make the decision as if it were entirely ours.
Discernment also draws on our instincts. When we have deep experience in a situation, we may sense that something is off before we can immediately explain why. That feeling can be useful, but it’s a signal to investigate, not proof that our conclusion is correct. Ask what triggered the concern, then use the steps above to test it against the facts and context. Few decisions arrive with complete information, so some comfort with uncertainty is part of the skill, too.
Which LEAD step would be hardest to take when an answer is well written and confident?
When has an instinct told me something was off, and what did I do to check it?
Discernment requires us to question and verify information regularly, not just when something feels wrong. These best practices can help us use technology without handing over our judgment or responsibility:
Clarify Before You Ask: Before turning to a search or AI assistant, identify what you already know, what you're trying to determine, and any assumptions you may be bringing into the question. This gives you a clearer purpose for using the tool and makes it easier to recognize when an answer misses something important. In practice, this might mean jotting down the facts you already have or outlining a project goal before asking for help.
Protect Privacy and Confidentiality: Before entering confidential or sensitive information into a digital tool, know where that information may go and who may be able to access it. Use approved tools and follow your organization's privacy, confidentiality, and data-handling policies. Having access to information doesn't automatically mean it's appropriate to share or use it.
Review According to Risk: Technology can help us get things done faster, but we're still responsible for the final result. A summary for your own reference may need only a quick check, while a client communication or decision affecting someone's job deserves much closer review. The greater the consequences of getting something wrong, the more carefully we should verify key facts, figures, and sources.
Own the Decision: Technology may help process information or recommend a course of action, but responsibility for the final decision still belongs to you. If you approve or share something, you should be able to explain the reasoning in your own words and stand behind the result. As compliance officer Betty Yang put it at a 2026 industry conference, "the outcome remains yours."
Be Open About How Technology Is Used: When AI influences a decision that affects someone, be clear about the role it played. Shah Md of HEPMIL Media Group describes a framework called ERA: explain how AI is being used, reveal its role and limitations, and allow appeal so people can question an outcome. If AI helps draft something like a performance review, a person should still review, own, and be able to defend every judgment that affects the employee.
Prioritize Human Connection: Technology can make communication faster while also creating distance from the people affected by our decisions. Don't let dashboards, messages, or generated summaries become substitutes for understanding the people behind the information. Team members we rarely see in person are easier to overlook, so make intentional contact with them and invite them to share what they're experiencing. For sensitive conversations or decisions, choose a communication method that gives people a meaningful chance to ask questions, add context, and respond.
Technology has changed the pace of leadership and expanded the reach of our decisions. What hasn't changed is our responsibility to apply discernment, own the choices we make, act with integrity, and remain connected to the people affected by those choices. This month, pay attention to one place where technology influences your decisions. Before you act, take the time to question what you're given, consider what's missing, and make sure the final decision is one you can explain and confidently stand behind.
Elevate your understanding of Ethical Leadership in the AI Era by taking flight with the following resources. Use this opportunity to navigate, uncover, and expand the horizons of your leadership influence.
Discernment: The Most Important Skill in an AI World
7 Practices to Build Discernment as a Leader
AI accountability in action: Key takeaways for strengthening responsible AI governance
