Four years ago, I posted about the importance of scenario planning. A few weeks ago, I posted about the urgent need for leaders to step up and design the future. What I didn’t mention, because it seemed obvious, is that between these two posts, years have passed and the world has experienced the ubiquity of agentic AI. We all use it, and we’ve also seen how AI underscores the importance of thoughtful, far-horizon, human oversight.
It seems that AI should make scenario planning easier, as it does with so many things, but it also brings inherent risks. A popular complaint is that AI “makes us stupid.” That’s just stupid lazy overly simplistic. I take issue with the word “makes,” as if we are zombies with no personal agency. As with any other resource (e.g., gunpowder, machines, the internet), agentic AI has the power to amplify actions without judging them; ambition, laziness, entrepreneurship, goodwill, and old-fashioned greed all feed equally at this trough.
So, to get a bit meta, we need to be more discerning in regard to discernment. McKinsey cites the findings of a Wharton study suggesting that dependence on AI can weaken our ability to practice and develop critical thinking. Agreed. But what I find compelling is that the same research also points to an opportunity to offset this kind of cognitive surrender with old-fashioned cognitive bias. I’m only being slightly sarcastic here.
Here’s what the research indicates. CEO’s have always had to make critical decisions based on whatever information is available at the time. What’s changed is that now they can quickly access comprehensive data, analyzed and neatly packaged, thanks to their favorite AI agent.
Case in point: when my leadership team engaged in scenario planning back in 2019, our data was based on input from strategists, tenured professionals, and trusted agency consultants. It took months of research and required the commitment of our top people over many, many hours of dedicated planning. All those engaged in the process were fallible human beings with our own biases. But we knew that. We even talked about the pros and cons of our various, strongly felt, opinions.
Enter AI. Our proprietary agents are capable of crunching vast amounts of data to surface predictive trends and spotlight blind sides in a way we couldn’t have imagined back then. The input is amazing and scenario planning should be able to advance much more quickly. But there’s a catch. The AI is generally pretty confident in its recommendations. If there’s a bias or outright error, it may be inherently blind to its own blind spots. And unless human beings, with all their faults and defaults, are part of the process, AI could drive the bus off the proverbial cliff.
This is where a strong CEO can step into the gap. Authorize the predictive trend research and deploy AI in collecting and sorting the data necessary for long-term planning. And share that with a diverse collection of executive leaders. Invite opinions, suspicions, and intuitions. Push back. Demand validation from AI; respect the input of humans. And build a comprehensive plan, including a menu of if/then scenarios, that will allow your organization to quickly respond to both calamity and opportunity. Continuously update the plan. Course-correct as necessary. And keep your human hands on the wheel.