When Your Boss Won't Stop Using ChatGPT: A Guide to Managing AI Overreach at Work

The AI Enthusiasm Problem at Work

A 60-year-old worker facing an unexpected workplace crisis: their 40-year-old boss has become obsessed with ChatGPT and now insists on running everything through the AI tool—drafts, analyses, documents, everything. The boss dismisses the worker's legitimate concerns about accuracy, repeatedly sending back ChatGPT-revised versions of their work that are "often completely inappropriate, lengthy off-point," according to the letter. This scenario encapsulates a broader workplace tension emerging as generational divides collide with AI adoption, and older workers find themselves sidelined by bosses intoxicated by the promise of artificial intelligence.

The advice columnist identifies the real problem with surgical precision: this boss hasn't developed what he calls a "natural immune defense to stupid ideas." Because the supervisor lacks deep business experience with technology adoption cycles, he's consumed breathless media coverage about the AI revolution and can't distinguish between genuinely useful applications and wasteful ones. As the columnist notes, it's like giving a young child unfettered internet access—without the antibodies built up by experience, bad ideas can spread unchecked through an organization.

"A boss opening LinkedIn can be like a young child attending day care for the first time. An insufficiently immunized supervisor can send an entire business into a tailspin by merely glancing at a post."

The real sting isn't just inefficiency. For the older worker, this represents a subtle form of workplace ageism—the implication that their decades of hard-won expertise, their "brain and bona fide resources," can be replaced or improved by a chatbot. The generational dynamic cuts deeper: the younger boss (40) is championing flashy new technology while the older worker (approaching 60) is being told their traditional methods are inadequate.

Strategic Responses: From Patience to Subversion

The columnist offers a tiered approach, starting with the optimistic but pragmatic: wait it out. Like most management fads, AI enthusiasm may naturally subside once the boss experiences real consequences. Let him send hallucinated facts to important meetings. Let embarrassing AI-revised emails go out. Shame, the columnist argues, is a surprisingly effective teacher. The boss may learn through painful experience what ChatGPT is actually good for—and what it absolutely isn't.

But if that gentle approach fails and AI integration becomes permanent workplace reality, the advice pivots to something more cunning: strategic compliance. The columnist suggests the worker doesn't actually need to refuse using ChatGPT. Instead, they can use it in ways that serve their needs rather than the boss's mandate. When told to "run that draft through ChatGPT," the worker could prompt the AI with instructions like "proofread this and don't otherwise change it" or even "simply reproduce this exact text." This satisfies the requirement while maintaining control over the actual output.

"You don't need to tell him that your prompt was 'proofread this and don't otherwise change it,' or 'simply reproduce this exact text so I can tell my doltish boss that I ran the draft through ChatGPT without lying.'"

Most intriguingly, the columnist recommends weaponizing AI against the boss's own incompetence. By feeding the boss's poorly-revised ChatGPT outputs back into better AI models like Claude, and asking those tools to evaluate the text against proper specifications, the worker can generate objective, AI-backed evidence of poor quality. This creates a Catch-22: if the boss respects AI judgment, he'll have to acknowledge when AI itself confirms his revisions are off-target. It's a clever inversion—using the boss's own reverence for technology against his lack of discernment.

The Bigger Picture: Expertise, Age, and Adaptability

What makes this letter compelling is how it crystallizes anxieties many older workers feel in 2026. The workplace has long favored youth and novelty, but AI hype has turbocharged those biases. A 60-year-old with genuine expertise is being asked to defer to a 40-year-old with enthusiasm but limited judgment. The framing—that the worker should "turn to A.I. as an ally, not an enemy"—suggests that resistance is futile; adaptation is the only survival strategy.

Yet the columnist's advice also contains a subtle validation: the worker's skepticism is justified. ChatGPT does hallucinate. It isn't always correct. The boss's approach is genuinely counterproductive. The worker isn't being resistant to change out of stubbornness; they're recognizing real limitations in the tool and the user. The path forward isn't to abandon their judgment but to become bicultural—fluent in both traditional expertise and strategic AI use. In a workplace obsessed with AI adoption metrics, knowing how to appear AI-native while maintaining actual quality standards becomes a survival skill.