@eviraBCG surveyed 70 senior executives and found something that should worry every company deploying AI at scale: Half are already seeing human judgment erode inside their organizations. More than 60% expect the problem to become material within the next 3 to 5 years. Yet only 1 in 10 has an organized response. But I think we may be focusing on the wrong solution. One recommendation is to stop measuring only what employees produce and start evaluating how they arrived at the answer. I wouldn’t build around that. Because “show me how you got there” is probably one of the easiest things to fake in an AI native workplace. Soon, we won’t be evaluating someone’s reasoning process. We’ll be evaluating an AI generated explanation of how that person supposedly reached an AI assisted answer. The more interesting part of the BCG research is how it defines one of the skills most at risk: Judgment and decision making. You face a problem with incomplete information. You make a call. You observe what happens. You update your model of the world. The critical word is observe. Judgment isn’t built by producing answers. It’s built by making decisions, living with the consequences, discovering when you were wrong, and adjusting the next time. Repetition matters because the repetitions carry consequences. And THAT may be what AI is quietly removing. More than half of the leaders surveyed report declining ownership and accountability. Today there is an incredibly convenient escape hatch: “The AI suggested it.” That’s not primarily a measurement problem. It’s an accountability problem. And no new performance rating system is going to solve it. Companies shouldn’t respond by monitoring employees’ reasoning more aggressively. They should create environments where people can use AI extensively, but still have to own the final call. Use AI to research. Use AI to challenge assumptions. Use AI to model scenarios. Use AI to recommend. But eventually a human should still have to say: “This is my decision. Here’s why I’m making it.” And, critically, that same person should later see what happened. Was the decision right? What did they miss? What changed? What would they do differently next time? That’s the feedback loop that builds judgment. BCG’s own first mitigation strategy is “Set the Organizational Conditions.” I think that’s where the real answer lies. The biggest long term organizational risk from AI may not be that humans become less intelligent. It may be that humans become less accountable. And if we remove ownership of decisions, we may eventually discover that we also removed one of the primary mechanisms through which people develop judgment in the first place.

@eviraBCG surveyed 70 senior executives and found something that should worry every company deploying AI at scale: Half are already seeing human judgment erode inside their organizations. More than 60% expect the problem to become material within the next 3 to 5 years. Yet only 1 in 10 has an organized response. But I think we may be focusing on the wrong solution. One recommendation is to stop measuring only what employees produce and start evaluating how they arrived at the answer. I wouldn’t build around that. Because “show me how you got there” is probably one of the easiest things to fake in an AI native workplace. Soon, we won’t be evaluating someone’s reasoning process. We’ll be evaluating an AI generated explanation of how that person supposedly reached an AI assisted answer. The more interesting part of the BCG research is how it defines one of the skills most at risk: Judgment and decision making. You face a problem with incomplete information. You make a call. You observe what happens. You update your model of the world. The critical word is observe. Judgment isn’t built by producing answers. It’s built by making decisions, living with the consequences, discovering when you were wrong, and adjusting the next time. Repetition matters because the repetitions carry consequences. And THAT may be what AI is quietly removing. More than half of the leaders surveyed report declining ownership and accountability. Today there is an incredibly convenient escape hatch: “The AI suggested it.” That’s not primarily a measurement problem. It’s an accountability problem. And no new performance rating system is going to solve it. Companies shouldn’t respond by monitoring employees’ reasoning more aggressively. They should create environments where people can use AI extensively, but still have to own the final call. Use AI to research. Use AI to challenge assumptions. Use AI to model scenarios. Use AI to recommend. But eventually a human should still have to say: “This is my decision. Here’s why I’m making it.” And, critically, that same person should later see what happened. Was the decision right? What did they miss? What changed? What would they do differently next time? That’s the feedback loop that builds judgment. BCG’s own first mitigation strategy is “Set the Organizational Conditions.” I think that’s where the real answer lies. The biggest long term organizational risk from AI may not be that humans become less intelligent. It may be that humans become less accountable. And if we remove ownership of decisions, we may eventually discover that we also removed one of the primary mechanisms through which people develop judgment in the first place.