- 6 de October, 2026
- Posted by: Filipa Ferreira
- Category: Leadership
Artificial intelligence gives leaders faster access to information, analysis and recommendations. But that efficiency carries a risk: the gradual substitution of independent judgement by automated answers.
Original judgement is the human capacity to see beyond the dominant narrative and make choices under uncertainty. It depends on two capabilities: breadth of perception and independence of interpretation.
The danger is not that leaders use AI. It is that they stop forming a view before consulting it.
1 Recognise the shift from support to deference
Cognitive offloading is useful until it becomes habitual. Leaders begin by delegating analysis, then defer to outputs they do not fully understand and eventually converge around the same metrics, models and narratives.
AI should inform judgement, not replace its formation.
2 Build structured curiosity
Curiosity needs operating discipline. Under pressure, organisations default to familiar explanations and close decisions too early. Leaders can counter this tendency through four practices:
Collect from the edges. Seek weak signals, outliers, unusual sources and perspectives from adjacent fields.
Arrange disparate signals. Place observations side by side so that patterns can emerge before anomalies are averaged away.
Suspend premature closure. Protect time for alternatives before a model recommendation or dominant narrative becomes the accepted frame.
Trigger alternative interpretations. Convert reflection into explicit questions and decision rules that can challenge habitual responses.
More data does not guarantee better judgement. Variety and interpretation matter.
3 Institutionalise intentional dissent
AI outputs often carry an appearance of objectivity that makes disagreement seem irrational. Original judgement therefore requires organisations to make challenge legitimate and repeatable.
Recruit independent perspectives. Bring in people whose expertise or cognitive style differs from the dominant group.
Use structured challenge. Test assumptions through pre-mortems, opposing interpretations and explicit examination of how a decision could fail.
Sustain dialogue over time. Dissent becomes valuable when people can challenge one another without destroying trust.
Healthy dissent is not disruption. It is protection against manufactured consensus.
4 Lead beyond the age of deference
The choice is not between human judgement and machine intelligence. The leadership challenge is to combine the reach of AI with the capacity to question its conclusions.
Organisations that preserve curiosity, independent interpretation and disciplined dissent will be better equipped to identify inflection points and make decisions that competitors do not see.
The more persuasive the system, the more deliberate the leader must be.
AI can accelerate a decision. Leadership still has to originate it.
Adapted for PTCG from: Leonid Sudakov and Nathan Furr, AI Is Undermining Leaders’ Judgment. Here’s What to Do About It, Harvard Business Review, 19 August 2026.