AI and the Men Who Would Be King
When the Divine Bleeds
In John Huston’s 1975 movie The Man Who Would Be King, Sean Connery and Michael Caine play two British rogue ex-soldiers who enter a remote kingdom, exploit a myth, and briefly turn performance into the power to raid the treasure room. Connery’s Daniel Dravot is mistaken for a divine ruler; the illusion holds until he shows a drop of blood. Once the divine ruler is revealed to be a mere mortal, the people’s spell breaks. Sean Connery’s Dravot loses his head, while Caine’s Peachy Carnehan barely survives to tell the tale. (en.wikipedia.org)
That is the useful frame for today’s AI news: what happens when the public starts seeing a small drop of fallibility in AI and the executives behind the myth are discovered raiding the public treasure room.
In the News
The public-facing miracle is weakening. Gartner reported this month that 49% of U.S. consumers say generative AI has made content quality worse. That is a problem for an industry still selling AI as a quality multiplier. For many users, the day-to-day experience is becoming more synthetic content, more sameness, more noise, and less trust. (gartner.com)
The brand aura is weakening too. TechCrunch reported that 60% of U.S. consumers say seeing “AI” in brand messaging is a turnoff, while 86% do not fully trust AI-generated answers without checking original sources. “AI-powered” is starting to sound less like a promise and more like a warning label. (techcrunch.com)
Clutch added another version of the same story this week: 93% of U.S. consumers say it matters that brand communications feel like they came from a real person, and 55% view a brand less favorably once they can tell AI produced the creative. The resistance is not simply anti-technology. It is anti-substitution - especially where companies replace human texture with synthetic scale. (clutch.co)
Inside companies, the resistance is quieter but just as important. WalkMe’s 2026 digital adoption study found that 54% of workers bypassed AI tools and completed tasks manually at least once in the prior 30 days. The future of work does not arrive just because management buys the software. Workers still have to believe the tool improves the work. (walkme.com)
And outside the screen, AI is meeting local resistance. Gallup reported in May that 7 in 10 Americans oppose construction of AI data centers in their local area, while Axios reported this week that data centers have become one of the clearest faces of AI backlash, with water and energy concerns moving from technical footnote to local politics. (news.gallup.com)
What If the Miracles Become Ordinary?
The risk for AI may not be dramatic collapse. It may be demotion.
AI may remain useful, valuable, and widely adopted - but no longer sacred. It becomes software. It gets audited. It gets complained about. It gets turned off. It gets procured, measured, governed, and distrusted in the same ordinary way people distrust other corporate systems.
That would still be a major technological shift. But it would be a different cultural story from the one the industry has been selling.
The public does not need to conclude that AI is useless. It only needs to conclude that AI is not divine.
The Men Raiding the Treasure
This is where the Man Who Would Be King metaphor sharpens.
The problem in the story is not only that Sean Connery is mistaken for a god. It is that the godhood opens the treasure room. The miracle creates authority; the authority creates access; the access creates extraction.
AI has its own version of this. While the public sees the idol up front - the chatbot, the demo, the keynote, the promise of intelligence - the earthly representatives behind it are collecting the spoils: venture capital, market IPOs, government protections, cloud commitments, enterprise budgets, user data, and public resources.
Some of that is legitimate business. Real tools cost money. Real infrastructure requires investment. Real productivity should be rewarded.
But the moral risk begins when miracle language is used to blunt ordinary scrutiny. The FTC’s Air AI settlement is the plainest version: the agency said the company and its owners would be banned from marketing business opportunities after allegations that they misled entrepreneurs and small businesses with AI-related claims about growth, earnings potential, and refund guarantees. (ftc.gov)
That is not the whole AI industry. But it is a warning label for the era.
Every gold rush attracts builders. It also attracts people selling maps to gold that may not exist.
Orthogonal Take
The question now is who survives the loss of mystique.
Probably not the brands that replaced human credibility with synthetic output. Probably not the vendors that treated “AI-powered” as a substitute for product quality. Probably not the executives who treated adoption as proof of value while the treasure room was still open.
The survivors will be the ones that can endure boring questions: Does it work? Is it accurate? Does anyone trust it? Can users opt out? Does it make the job better? What does it cost? Who benefits? Who pays? Who is accountable when it fails?
The god does not have to perish.
It only has to show a small drop of blood for people to stop believing that it’s infallibly divine - and to start asking which mortal men raided the public treasury.