No, it's not even really AI at all, it's deep learning. But... that's still bad and the sad thing is people are volunteering, nay queuing up to help it learn their emotions/ facial expressions etc etc. There are already classrooms for "disturbed" children being monitored by "AI" looking out for trouble and nipping it in the bud before it happens based on facial expressions etc. it's not hard to foresee a time when you will be judged before you do anything. For instance entering a shop you'll be stopped at the door because the AI has decided you're likely to be a shop lifter. where will it end?
Did you know you can use Chat GP (think it's called that) to write programs simply by describing what you want the program to do. Apparently it's pretty good at it too. The future skilled people will be the ones who can instruct the AI most succinctly. not good at language? You'll be screwed. The people who will be good at this are probably the type of people who are rather robotic and cold themselves.
Can anyone get it to do a vector graphic image of a late bay on a tow truck with a plain white background, with a bay that has correct bumper etc, and in decent enough resolution to use for a sticker? I've failed.
Authors are already up in arms as people are getting AI to write books in the style of other writers. And how many people have got degrees because AI did the work for them.
Its always been possible to get someone else to write your dissertation, copy someone elses work, get an essay written for you by paying someone etc. Agree this makes it easier but getting a degree is much more than just submitting written work, and good academics will spot fakers a mile away. Chap on my course submitted someone elses work for a single paper and got found out - didn't match his tone/ usual way of writing. Got kicked out. yep, all its doing is searching for things that already exist and doing its own version of them - that's why you get early indicators on a late bay etc - just like if you google search late bay you'll see all sorts of busses. it doesn't "know" which is right so it merges/ combines attributes from all the examples it can find, not all of which are correct. Ditto the point about university work - it may put together a convincing essay but it may not actually get things right - its only as good as its source and as that is the internet then it may not be correct!
Indeed. AI doesn't "know" anything. Unfortunately, brainless politicians have now cottened on to it, and you get the usual hysterical nonsense. Recent case of the AI expert in the US - one of the AI pioneers - who has been "disinvited" to the forthcoming AI conference. His crime was pointing out it wasn't half as clever as the jabbering classes think it is.
Learning how to write the "AI" prompt so that the picture looks like the one the customer wants and doesnt have something like Donald Campbells Tomato Sauce" written clearly on the label of the soup tins. It takes a lot of attempts, and in a few years as the systems plagiarize each others rubbish it will become less and less useful. Especially if a few clever people start deliberately mislabelling dodgy pictures as something else.. A lot of illustrators were already producing derivative works using nothing more than their brains and Photoshop..
And when they do they will be rolling around laughing at their colleagues attempts to do Arnie impersonations...