๐๐ก๐ ๐๐ง๐-๐๐๐ฒ ๐๐จ๐จ๐ซ
7/8/20261 min read


AI won't seize control. You'll hand it over one reasonable choice at a timeโฆand the door will lock behind you.
I argued recently that AI speed makes human judgment look like delay. Letโs follow that same road all the way to its logical end. The danger is not a machine grabbing control. It is a workforce that slowly loses the ability to take that control back.
Nobody decides to become dependent. You simply keep accepting the faster answer, and ๐๐๐๐ก ๐ก๐๐ง๐๐จ๐๐ ๐ฅ๐จ๐จ๐ค๐ฌ ๐ฌ๐๐ง๐ฌ๐ข๐๐ฅ๐ ๐จ๐ง ๐ข๐ญ๐ฌ ๐จ๐ฐ๐ง: quicker, cheaper, usually right. Whatโs not to like? The trap is never the first decision. It is the thousandth.
Here is the loop that quietly locks the door.
Microsoft and Carnegie Mellon studied 319 professionals using AI at work, and found that the more they trusted the tool the less they thought for themselves, while the more they trusted their own skill the more they did. So ๐๐ฏ๐๐ซ๐ฒ ๐ก๐๐ง๐๐จ๐๐ ๐ช๐ฎ๐ข๐๐ญ๐ฅ๐ฒ ๐๐ซ๐๐ข๐ง๐ฌ ๐ ๐ฅ๐ข๐ญ๐ญ๐ฅ๐ ๐๐จ๐ง๐๐ข๐๐๐ง๐๐ in your own judgment, while the machine never doubts itself for a second.
The risk is how it can end. You keep losing more arguments with the system, not because it is right, but because it is certain and you no longer are.
Past a certain point, doing it yourself feels riskier than simply deferring again, so the โoverride AIโ button still exists but fewer and fewer people can or dare actually press it. This is something worse than agency left unused: it is ๐๐ ๐๐ง๐๐ฒ ๐ฒ๐จ๐ฎ ๐๐๐ง ๐ง๐จ ๐ฅ๐จ๐ง๐ ๐๐ซ ๐ฉ๐๐ซ๐๐จ๐ซ๐ฆ.
But none of this is fate, and the door is still openโฆ๐๐ ๐ฎ๐ค๐ช ๐๐๐ฉ ๐ฉ๐ค๐๐๐ฎ.
Aviation already faced this. After automation was linked to fatal crashes, regulators urged airlines to switch the autopilot off and let pilots hand-fly on purpose, not because the machines fail often, but to keep humans capable for the moment they do.
Treat ๐ก๐ฎ๐ฆ๐๐ง ๐ฃ๐ฎ๐๐ ๐ฆ๐๐ง๐ญ the same way: ๐ง๐จ๐ญ ๐ ๐๐จ๐ฌ๐ญ ๐ญ๐จ ๐จ๐ฉ๐ญ๐ข๐ฆ๐ข๐ณ๐ ๐๐ฐ๐๐ฒ, but a strategic reserve you cultivate and maintain on purpose. This ensures you still have people who remember โhowโ and you can remain ๐ฆ๐๐ฌ๐ญ๐๐ซ ๐จ๐ ๐ฒ๐จ๐ฎ๐ซ ๐๐๐ฌ๐ญ๐ข๐ง๐ฒ.
Which human capabilities is your organization keeping in practice deliberately, and which are you quietly letting rust?
#AI #FutureOfWork #HumanInCharge #AIStrategy #HumanJudgment
Contact
bruno.gentil@sherpaconsultingasia.com
ยฉ 2026. All rights reserved.
