I think OP is reacting too much to the implied threat of chatGPT. It's not so good as it seems. If it does the job in 1:10 the time and still makes errors, then

I think OP is reacting too much to the implied threat of chatGPT. It's not so good as it seems. If it does the job in 1:10 the time and still makes errors, then it might speed up productivity by a mere 2x because manual review is hard. In my experience there is no task where chatGPT is 100% reliable, if it is high stakes it needs review. And chatGPT might not be best - for almost any task there is a specialised model that does it better. Humans make different kinds of errors anyway, we are GPT-N level for a large value of N - so a combination of AI+human might give best results.