Ever wonder why smart groups often make poor decisions?
Businesses and institutions often slip into ‘Groupthink? From casual groups to formal teams, even when aware of their tendency toward confirmation bias, they naturally favour opinions aligning with prevailing views.
At its worst, Groupthink means ignoring opportunities to consider differing opinions and data and dismissing them when they are presented. This usually leads to choices that with the benefit of hindsight are clearly stupid. Think of it as everyone boarding the wrong train because no one dared to question the destination.
Alignment, that often used management cliche however, is essential for optimal performance. Everyone on the team should clearly understand the direction they’re heading and why that direction matters. To extend the metaphor, everyone on the train knows where it is going, what their role is, what they need to do on the journey to arrive at the declared destination.
True alignment happens when all opinions, information, and data have been carefully considered, weighed, and distilled into a clear consensus. The best choice is obvious, and everyone either fully supports it or at least understands it as the optimal route forward. The strategic challenge is ensuring the destination to which all are aligned, is the optimal choice given the strategic, competitive, and regulatory context.
Are ‘Groupthink’ and ‘Alignment’ synonyms? Or just two sides of the same coin?
Groupthink: Bad. Alignment: Good.
Both can suffer from confirmation bias, even when teams consciously try to avoid it. Alignment can become especially dangerous if unchecked confirmation bias sneaks in.
Many strategies exist to ensure the best choices emerge from challenging decisions. Employing a Devil’s Advocate is one approach to removing any pre-existing bias. It includes techniques like ‘red teaming’, or involving independent external experts for objective interrogation.
Chat GPT 4.5 recently landed in my account with its ‘Deep Research’ capability.
This marks a genuine leap forward for AI.
Earlier models like Chat 3.5 already enabled the asking reflective questions like, “What have I missed?” or “What should I be asking?” Although useful, these prompts typically delivered limited responses.
Chat 4.5 with Deep Research elevates the Devil’s Advocate approach to an entirely new level. It deeply interrogates the topic, reasoning through provided prompts and resources to deliver nuanced, sophisticated, and profoundly useful insights.
This capability changes the game for management teams, provided their commitment to a particular viewpoint doesn’t block genuine consideration of alternatives.
l remember Bill Shorten’s absurd 2012 ‘blind support’ gaffe when asked for a response to PM Julia Gillard’s removal of Peter Slipper as Speaker. He said, “I haven’t seen what she said, but I support whatever it is that she said.” While Shorten understandably wanted to avoid contradicting the PM, his words perfectly illustrate how blindly following a position without any questioning is just dumb. He was however, perfectly aligned with the PM, useful when climbing slippery leadership poles.
