Artificial Intelligence Can Seriously Damage Your Career
A Public Health Warning
As businesses grow, what worked for a startup tends to not be what’s needed for the next phase for an enterprise that has become established and proven its profitability
The name of the game shifts to growth from surviving the startup phase
I’ve been working with a small, but highly profitable company (let’s call it Company Q), helping to get them organized for their next phase of growth. Part of my work involves putting in place checks and balances, and controls and guardrails (known in the finance business as internal controls) so that things don’t go wrong in the normal course of events
Of course a bad actor who is determined to do harm will find a way to circumvent the controls, but that’s a different issue
One of Company Q’s trusted and long serving employees (let’s call him Steve) who has been with the business since it started made a bad call
Steve had previously asked the owner for permission to do something out of the ordinary with the payroll and the owner reluctantly agreed. A couple of months later, Steve made the same request which again was reluctantly approved. I asked the owner if he was aware of any problems in Steve’s personal life which might have led to these requests, but he was not aware of anything
The same unusual payroll matter popped up again, but this time without any discussion with the owner, who was very angry and clearly felt the trust he had placed in Steve had been betrayed. The owner mentioned that he was worried that if Steve had done this without permission, what else might be going on?
My advice was “let’s suspend Steve on full pay while we do a quick investigation into the areas of concern”. So, a suspension letter was issued, clearly stating the cause for the suspension
36 hours later no further irregularities were apparent and on the face of it, this was an isolated transgression. The owner and I decided to invite Steve\ to discuss the issue and share his version of events. Monday at 9:00am was the appointed time
When we set this meeting up, the owner was 50-50 on whether to keep Steve on board or let him go. Its tough to let someone go who had worked so hard to build the business
At 6:30am on Monday, the owner received a text message declining to meet
Here’s the text message Steve sent to Company Q’s owner:
I recognize the importance of being fully informed to address the current situation appropriately, given the circumstances surrounding my formal notice of suspension and complaint.
Could you please send me the details of the complaint before we speak?
We will need to reschedule our conversation as I will need to review what is provided prior to speaking. We can utilize Google Meet to maintain privacy and ensure a comfortable and structured conversation.
My email is Steve@Iamsomewherepanicking.com
Thank you for your understanding.
At 8:30 when I got to the office for the now cancelled meeting the owner was furious
“That’s it! Steve can’t stay. I can’t handle any more of this b*l$h!t” he exclaimed
The decision was made
Things moved fast; Steve’s access to the company bank account and payroll system was blocked, his email account was shut down and the locks to the office were changed
Apart from assuaging the owner’s rage, the goal was to protect the company from any fraud that we may not have already discovered to ensure that no damage could be inflicted on the business by Steve going forward
Steve was formerly terminated later that same day
After the dust settled, I reflected on the events
Following the suspension, it seems clear that Steve was worried about his future with the firm and he understood how upset the owner was
He turned to ChatGPT or some other AI agent for advice, and was provided with the text message that he used to blow up his one chance to save his job
Steve blindly used the AI response without running it through his own human intelligence system, which we normally call a brain. The complaint the AI agent was requesting information about was clearly and unambiguously stated in Steve’s suspension letter
Unfortunately, Steve didn’t have the presence of mind to share that with the AI agent
Steve (and I fear many others) really don’t understand that any answer from an AI large language model should be used with caution, and very often need a human touch
He failed to grasp one of the key issues associated with AI, namely that it’s a tool to help with a problem, rather than the whole answer itself
It’s very clear that results from AI agents need to be verified before being trusted (thanks to President Reagan for that prescient advice)
sadly, people who haven’t grasped AI’s current limitations are being emboldened - and in this case with disastrous results
Steve is now on the job market without a favorable reference from his employer of almost 5 years
“You can’t fix stupid” is an old saying
In 2025, we might be seeing the emergence of a complimentary idiom
“Artificial Intelligence + Natural Stupidity = Potential Disaster”