AI-loving boss insists that employee uses Chat-GPT to assemble company's monthly report, employee complies and submits absolute gibberish to upper management: 'The result was entirely and completely useless'

Author image for Ben Weiss
Ben Weiss • Writer and Editor
Dec 6, 2025 4:00 PM EST

It comes as no surprise that far too many employers out there are charmed by the so-called efficiency of AI in the workplace.

What's especially humorous about this concept is that these are the same managers who tend to be slow to embrace change. They didn't like the concept of fully remote work, and they don't enjoy more contemporary ideas like the four-day work week. However, when it comes to using AI and Chat-GPT in particular, some bosses have rushed too quickly to embrace these tools simply because they are seemingly cost-efficient.

Once again, what these employers are sacrificing is accuracy. We've seen this kind of sacrifice manifest itself in the workplace time and time again. Whether it's consolidating positions to save costs or eliminating certain elements of a due diligence process to maximize speed, a human attention to detail tends to be on the chopping block over and over again.

That's exactly what happened in this workplace environment, in which this author was encouraged to use Chat-GPT over their own tried and true methods of analysis for an upcoming monthly report. Despite numerous attempts on the part of the author to explain why this was not a good idea, the boss refused to budge. So the employee complied, albeit maliciously…

Add as a preferred
source on Google
Advertisement

  • When the employee submitted the monthly report to their boss, it was full of Chat-GPT-ridden errors because the tool was simply not advanced enough to properly analyze the company's data. The author expected the boss to catch these errors before presenting the report to a room full of coworkers and colleagues… but alas, it seems that this boss valued AI just a bit too much.

  • "Boss: You MUST use Chat-GPT! Me: OK!"

    Please forgive me, English isn't my first language. My boss is on the GPT hype train, you would think he's got stock in OpenAl with how much he uses it. My boss is also convinced that GPT can do no wrong. GPT is always right, even when it's not.
    via beerbellybegone
  • We should change reality to accommodate what GPT said, even if it goes against best practices or just doesn't make sense.
    via beerbellybegone
  • Whatever, more than it's worth for me to argue, why should I care?
    via beerbellybegone
  • The latest stupidity is he told me to use GPT to run an analysis on something that I generally do manually. Instead of objecting, I said sure, why not. The thing is that the source material I'm analyzing is EXTREMELY difficult
    via beerbellybegone
  • for GPT to parse. I uploaded just the first of about a dozen interlinked documents, and GPT was spitting out gibberish. Three documents in, GPT was hallucinating answers that were nowhere even close to what the results should be. After all the documents were uploaded, GPT
    via beerbellybegone
  • was crying for mama. The result was entirely, and completely useless. Garbage would have been a kind description.
    via beerbellybegone
  • I packaged up this rancid pile of discombobulated nonsense, used the standard monthly report template, and sent it off to boss- man. He *tried * presenting it at
    via beerbellybegone
  • the monthly meeting with the directors, but there's no way in the world he'd be able to pass that turd off as anything else but what it was - a turd.
    via beerbellybegone
  • I'll know in about 3.5 weeks if he wants me to use GPT on next month's report.
    via beerbellybegone
  • A young confident businessman in a suit gives a presentation to his colleagues in a conference room.
    via Unsplash+ (Getty Images)

    Image is representative only and does not depict the actual subjects of the story.

  • hacktheself You should bring the operational security risks of this policy up the chain, particularly if your boss has you uploading company sensitive data to that untrustable external company.
    via beerbellybegone
  • potatofriend26 > Please forgive me, English isn't my first language. > I packaged up this rancid pile of discombobulated nonsense lol
    via beerbellybegone
  • Thankyouhappy "GPT was hallucinating answers", not enough people in management understand that this is the achilles heel of this emerging tool.
    via beerbellybegone
  • Complete-Emergency99 Al is great!!/S I showed a co-worker a classic old website that he didnt knew existed. Then i got curious of for how long said website has existed. My guess was late 90's/early 00's. Googles own Al-assisted search result claimed that such webpage didn't even exist. Despite me looking at it 1,5 minutes earlier.
    via beerbellybegone
  • Ok-Description-4640 I work in 0365 and have basic familiarity with powershell and I'm extremely dubious of Al services. I finally broke down and tried to do an extremely simple task, asking Copilot to write a PS script to take a csv file and update the job title and manager in our AD for each user from the text file. It put out 30 lines of code that didn't work. I spent a day and a half trying to make it work, then said eff it and wrote six lines of code myself that worked. I'd still like to deb
    via beerbellybegone
  • copenhagen_bram > Three documents in, GPT was hallucinating answers that were nowhere even close to what the results should be. After all the documents were uploaded, GPT was crying for mama. Your English is fine :D
    via beerbellybegone
  • Philodendron69 "GPT was crying for mama" your English is EXCELLENT
    via beerbellybegone
  • tynorex The thing is, the more you understand a topic, the more you realize how unreliable Al is right now. Al is really good at sounding human, I think of it as like my smart friend who is generally knowledgeable, but who gets specifics mixed up often. Is it generally right? sure, but once you get to a detail level, it often is wrong. Google has been laughably bad, where I will see their Al writeup (I often have to double check state laws) and then I will see the summary from the State website
    via beerbellybegone
  • Flimsy-Bluejay-8052 Set a parameter in your bosses GPT that always sways the conversation in slight, reasonable ways, to allude that data analysts need to be paid more.
    via beerbellybegone
  • HumanTheTree Boss: well the ai generated report was pretty bad, BUT ai is always learning, so the next one is going to be better!
    via beerbellybegone

Want More? Follow Us and Add Us as a Preferred Source on Google.

Tags

workplace-stories work stories toxic-workplace job work Horrible Bosses company workplace malicious compliance toxic-work-environment
Scroll Down For The Next Article