Blaming the Tool
A friend told me about someone at their company who got caught using ChatGPT to answer client questions during a live meeting - copying the questions in, reading the responses back, right there in front of the client. They'd been in the role long enough that nobody had been checking their work too closely.
The answers were sometimes wrong, sometimes not useful to the client, and always obviously not coming from someone who understood what was being discussed. This person didn't have the knowledge to participate in the meeting they were sitting in, so they had an AI do it for them in real time.
Management got involved, and the AI usage became the focal point.
This person's inability to do the job wasn't new, though. Their coworkers had raised concerns for months. The feedback either didn't reach the right people or it did and nothing happened. The inability to do the work was there before ChatGPT existed. The company just didn't treat it as a problem worth solving until the tool showed up.
In my experience, this is how it usually goes: a company tolerates mediocre work for a long time because the output looks close enough. Nobody's really checking whether the answers are right - they're checking whether the answers exist. Then someone finds out the responses are coming from an AI, and the urgency shows up.
Sending unverified AI output to clients is a legitimate problem. If you don't understand the material, you can't tell when the AI gets it wrong, and that's a real risk. But the clients were already getting bad information before the AI was involved - the only thing that changed is how it was being generated.
A competent person using AI to draft or double-check their work isn't doing anything wrong - that's just a tool. But when AI becomes the entire basis for someone's output, when it's replacing expertise instead of supporting it, the actual problem is a hiring problem, or a management problem, or a performance review problem that nobody addressed.
If this person had been giving the same wrong answers by hand - just making things up or guessing from a meeting they hadn't prepared for - would anyone have noticed? If not, the company has a much bigger problem than which tools people are using.