With AI you can feel that you have gained super powers as a leader. You have a companion that is available to you 24/7, always welcoming you, always energetic and able to provide you with various perspectives on a given matter – built upon the collective knowledge of all information posted by all human beings. Should you use AI as support for decision-making? Absolutely, but you must understand its shortcomings as well as how your own bias plays in.
Indecisiveness is often seen as a weakness, especially if you have a managerial position. As humans we are used to the narrative of true leaders as those who have good communication skills, can make decisions as no one else, attract followers and inspire those to follow the path laid out by these decisions. But honestly, how often are you in doubt? Full disclosure: I am constantly in doubt and that concerns me, as I am paid to know and paid to provide some good advice for my customers. Self-confidence is king for a leader because the confidence spreads and provides comfort in the organisation. However, confidence is often misinterpreted as competence. You simply look smarter just because you have self-confidence, despite your flaws in competence. At least that is how it is for males; for females the cultural understanding might lead to a different perspective, as self-confident women can be seen as simply being "too big for her boots". But that is a different story.
So how do leaders deal with this doubt?
Many leaders use coaches who in a neutral and confident manner help them reflect and see new perspectives, and by that come to a decision. But a coach is not always available, and a coach costs money, so AI can be seen as a great alternative – which it is, as long as you know its limitations.
One of the first challenges you must know with AI is that in general it is guessing when it does not know. If information is not available, either online or you providing it, your agent will make a guess but not disclose it to you. So when you are using AI for decision-making, you must have discipline and describe the full context the best you can. Are there for example different perspectives, different opinions etc.? Yet this is not enough, because you cannot share more than what you are aware of. You have tacit knowledge, which only will come out when triggered. So in addition to providing the full context, you must also interact with your agent in a way that helps you reflect and realise things you did not think of at first. Ask your agent to challenge your assumptions and see the challenge from different angles. Even basic and obvious questions still can help remind yourself about unspoken things that make a difference.
Another important matter for using AI-based decision-making is that even though it is designed for strengthening your decision-making capabilities, there is a risk it ends up making it weaker. A field experiment from Harvard University explored evaluations of submissions to an MIT global social innovation challenge. Three conditions were tested: 1) human-only evaluations; 2) AI evaluations with a narrative explanation of the decision-making; and 3) AI evaluations with no further context. The evaluations were compared against four experts affiliated with the innovation challenge.
As shared in Harvard Business Review: What they found was striking: When the AI tool encouraged human evaluators to reject submissions that the expert panel approved, the humans largely went along with those recommendations, filtering out high-potential innovations. What's more, they were less likely to override an incorrect recommendation from the AI when offered a narrative explanation. Instead of improving the evaluators' judgement, the narrative explanation made it worse.
You are more likely to follow a wrong recommendation from AI, simply because it sounds convincing – or, you could say, it comes with great confidence. The more you use AI as a decision-making tool, the risk is you will lose fundamental judgement capabilities. You will lose the unique human capacity to see beyond the dominant narrative, and by that stray away from making choices that represent your values when being uncertain.
So to narrow it down: AI gives many advantages for leaders to help make good decisions, but see it only as one of the tools in your toolbox. Have an awareness of providing enough information to your agent, request the agent to help you see things from different perspectives, and in the end provide critical thinking to the recommendations and suggestions provided to you. In many ways it is like a pocket calculator: if you constantly use the calculator for even the simplest calculations, you will end up losing the ability to calculate yourself. Along with AI-based decision-making, strengthen your personal judgement capabilities and put attention to your gut feeling.
