Like everybody else, I've been playing with ChatGPT lately.
I'm happy to report that, thus far, AI currently poses no threat to the magical arts.
The following is the response I got when asking, "How does a magician tear and restore a newspaper"?

'A magician tears a newspaper into small pieces and then magically restores it to its original state. This trick is often achieved through a combination of misdirection and sleight of
hand. The magician might conceal a pre-cut newspaper or use special techniques to restore the torn pieces, such as using clear tape or glue to reattach them. The key to successfully
performing this trick is to distract the audience's attention while the restore is taking place, so that they do not notice the magician's movements.'

Most everything can be found online or in a public library. Try "How to force a card". I think I just learned a new method.

Err, with ChatGPT. Came back with a slew of forces!

Having worked in the IT-industry for quite a number of years, I can tell you how much better AI became over the past years: There have been massive improvements! This progress flies under the radar, since most people discard AI because they only think of what's called "strong AI": An AI you can have a meaningful conversation on Byron with and that can dance a Rumba and make a coffee while at the same time being able to write a poem about ... That is not where research has gone in past years, it was occupied with what is called "weak AI", AI-applications that are programmed to do a certain task and do it well. IBMs Deep-Blue chess-machine is a weak AI, the software your call-centre-agent uses to determine what you will want from him/her even before the phone is picked up is a weak AI - and so is the software that drives your fancy car in autonomous mode. Don't be fooled: Weak AI can be very, very impressive - like ChatGPT is, which is a weak AI combined with a massive (as in: humongous!) dataset.
If AI will be a "threat" to magicians, depends entirely on whether or not someone takes the time to feed a weak AI with good, tangible, specific data about magic or not. Once there is a "ChatGPT - Magic Edition", we may be in trouble.
Then again: There are loads of books on magic in public libraries (in Germany, at least) and still not everyone out there is a magician. So, maybe it's not so much a question of whether a "ChatGPT - Magic Edition" replaces us, but how _we_ use it.
My two (Euro-)cents ...

Much appreciated!