Totally get where you're coming from. It’s definitely going to shake things up — some roles will disappear, but new ones are already popping up too. I think the challenge is retraining fast enough. Not everyone wants to become a “prompt engineer,” but AI isn’t going away, so we kind of have to adapt or get left behind.
I was a hair too young to attend Woodstock back in 1969. Though I dislike AI-generated posts like the above, it was kind of funny to see what an AI-generated response would be to a question about the threat of AI, itself.
I use AI and I appreciate the fact that it does save me a lot of time - for research purposes and 1st draft content. A couple of points to consider here though. While you touched on it in your point #1 I believe that more emphasis is needed for empathy and insight. Those are two very important qualities that AI just does not have - yet. Also, AI lies. When you request research and content suggestions AI scrapes the net for related results. Its output has a lot of material that does relate but the fact is a lot of content on the web is just false. And though some of that content might be related in subject if you use it you don't provide quality AND you eventually damage your own reputation. All AI content requires review, edit, verification, and revision - by a real person - so it not only provides quality, factual content but also shows that human touch, along with delivery in the publisher's voice.
I am told that there are LLM's out there (I believe there is one called Perplexity) that offer links to the sources of their research answers. I have a paid version of ChatGPT and it occasionally shows me newer capabilities which include citation links, as well. So, hopefully, that will make fact-checking a bit easier.
The one thing I like about Chatgpr is that you can work with someone and not have to worry about their ego getting in the way.
I've been relying on Grok more frequently, and it's proven to be an excellent tool for refining scripts and code. Recently, I was uncertain about some prepared statements, but after running them through Grok, the results were spot-on. Even this response was polished with Grok's help.
There are a number of engineers in my company who are using an LLM (or multiple LLM's) to write emails and other communications for them. As a sales and marketing guy who has been writing his own snail mails, emails, and other comm's for decades, I just write and post from the hip. I am using ChatGPT more of as a senior advisor that I ask questions on any topic that I can think of than as a writing assistant.
There is constant improvement. I've taken to include requests in my prompt for the sources of any claims. It's working with ChatGPT for me. It does help but the information must be verified.