{
  "id": 1665086,
  "title": "Why our leadership isn’t ready for AI (Part 6)",
  "url": "https://urgent.news/2026/08/18/why-our-leadership-isnt-ready-for-ai-part-6",
  "topic": "ai",
  "section": "AI",
  "published": "2026-08-18T06:56:00.000Z",
  "source": {
    "name": "Bangkok Post Business",
    "slug": "bangkok-post-business",
    "url": "https://www.bangkokpost.com/business/general/3303767/why-our-leadership-isnt-ready-for-ai-part-6"
  },
  "original_language": "en",
  "account": "Artificial intelligence doesn’t eliminate the need for human judgement, but rather it can unintentionally remove it from the decision-making process. Some leaders worry that their teams become overly reliant on AI outputs without critically assessing them. However, the reality is that judgement might not be disappearing, but rather being systematically removed from the system. When speed is prioritized, AI-generated suggestions are often treated as final answers rather than drafts that require verification. In situations where questioning the AI takes longer and is penalized, people tend to accept the AI's recommendations without further scrutiny. An example of this dynamic is when a report arrives with an AI-generated recommendation attached to it. If the deadline is imminent, the team lead who challenges the figure and demands a second opinion is often perceived as slowing down the process, while the one who forwards it promptly appears more efficient. Over time, organizations inadvertently train their most capable individuals to stop questioning the AI's outputs.\n\nJudgement requires friction, meaning it necessitates a brief pause to analyze the situation and question whether the AI's output aligns with other available information. Additionally, permission is required to slow down even by a few minutes, especially when working under tight deadlines. Permission to express disagreement with the AI's output in writing, without it being perceived as a lack of confidence in the tool or oneself, is also crucial. Furthermore, expressing uncertainty should be viewed as a valuable signal rather than something that needs to be smoothed over. If these conditions are not met, oversight becomes a passive approval process where the reviewer simply signs off without actually verifying the AI's recommendations. This kind of passive approval erodes quality gradually, long before anyone notices it. Ultimately, a flawed recommendation often goes unnoticed until it causes significant damage, at which point everyone wonders how it was approved. The reason for this is almost always the same: the system made approval effortless while making disagreement costly.\n\nOrganizations that maintain sharp human judgement do not simply preach critical thinking; instead, they design and incorporate review moments that explicitly ask, \"What would you change about this and why?\" Instead of relying on straightforward \"approve\" or \"reject\" clicks, they protect time in their calendars for a thorough second look at critical decisions and treat that time as non-negotiable. Moreover, they recognize and celebrate individuals who catch the AI's mistakes as highly as those who ship projects quickly. This approach may initially seem counterintuitive, as the promise of AI was to reduce friction and eliminate the need for second-guessing. However, deliberately reintroducing a small amount of friction precisely where it matters most can be beneficial. For instance, a ten-minute pause before making a decision that impacts customer relationships, hiring, or significant financial commitments is not the same as a ten-minute pause before approving a routine internal memo. The organizations that successfully implement this strategy do not become slower overall; rather, they are deliberately slower in the specific instances where speed has quietly been compromising their judgment while remaining as fast in all other areas. This deliberate inclusion of friction where judgement is crucial prevents the slow, unnoticed erosion of critical thinking that can occur over time, as each moment of avoiding disagreement gradually diminishes the organization's ability to question and assess AI outputs effectively.",
  "summary": "Many leaders worry their teams are becoming too dependent on artificial intelligence: accepting outputs without question, losing the instinct to challenge, nodding along to a recommendation nobody actually questioned.",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}