ChatGPT is not what people think it is.
There has been an experiment where somebody told ChatGPT to summarize the content in a link.
What ChatGPT did was make information up based on the keywords that are in the URL of the link.
Researchers put numbers in the URL to validate their initial suspicions, so there was no way for ChatGPT to "guess" the page's content.
What result did you think ChatGPT returned?
That's right, "I am an AI robot."
You know the spiel: I am an AI software etc.
The bottom line is that ChatGPT is a language model, meaning it takes bits and pieces of information and slaps them together into a believable format.
In many cases, this works if the information is known and there is lots of data on the internet.
But sometimes, the information is so off-base that ChatGPT can be accused of "hallucinating" or making stuff up.
I am sure that AI writing will evolve and get better, but we're not there yet.
You are, at this point, taking a huge risk that whatever output you get from ChatGPT can be confidently posted online.
It has to be edited and fact-checked.
Also, there's the human engagement factor because, at the end of the day, regardless of how you produce your content, SEO success requires delivering on the search intent behind the keyword the reader of the content has.
I hope you understand how this works because you will get great results if you do.
It's an open question as to what the future of online publishing will be.
If search switches over to a chat format where the search engine gives you the answer, there will be fewer incentives for websites because people aren't going to look at the sources anymore.
They will take the answer and leave, so this means they aren't going to click on your website or click on your ads.
That's the real challenge that ChatGPT poses.
It's not so much about its technical capabilities today but what it means in terms of its introduction to the world of a chat-based search interface.
You can bet that things are going to be interesting.