Right. ChatGPT draws on the internet for information. But the internet is a mix of true facts and BS. It takes a discerning human mind to know what is true and false, and sometimes that is hard to do.
Another AI expert warns of itās dangers. An interesting point he makes: AI code is grown organically within supercomputers rather than being āprogrammedā by humans. We donāt really understand it or how it works in any logical way. Itās like a growing organism and itās intelligence will soon far surpass our own. This is what Kurzweil calls the Singularity and Kurzwel thinks itās great. But we donāt really understand what we are unleashing on this world or where this will end up.
Yes and no in regards to AI mirroring the internetās facts and BS. The broad model is trained on internet data but many plug-ins have databases that have been specifically selected for specialized topics. They are trained on specific texts. Also, AI provides sources if asked on their primary model, allowing users to follow up and fact check what was written by AI.
Examples of models trained on specific topics:
Claude has just arrived in Europe and Iām loving it! It can transform my notes into text and it has a better Portuguese than Chatgpt4! yehaaaaaay!
@_Barry Yes, itās now a big problem for college professors because students are using AI to write papers vs. doing their own research and original thinking. I think it can lead to a kind of intellectual laziness if we use AI to do our thinking for us.
Late to the party, as usualā¦ Scrolling through this thread. Not clear to me from the IMDB Page if this movie (The Artifice Girl) is supposed to be fiction or not but I know for a fact AI has already been used successfully for this purpose, I saw a documentary about it.
This could be the death knell for many online websites, including online news publication and other websites, even Wikipedia.
I just finished grad school and there was a student who was doing this for a group project and it was obvious to the group mates (and maybe the professors). At least for my discipline, AIā¦ or at least the platform he was using for itā¦ wasnāt doing a good enough job. I suspect that will change with time though. For fun I had tried plugging in things to see if it could write a paper for me (I would never cheat though) and what it wrote sucked. It might work for a freshman undergrad though.
@_Barry Yeah, we now have to question everything we see online or in the media because AI ādeep fakeā technology is so good. Donāt be surprised if we soon see a video of the Dalai Lama praising the greatness of the Peopleās Republic of China. Or fake videos of various celebrities telling us who to vote for.
@Parhelion Yeah I think one can usually tell if a paper is written by AI, but a lot of students are trying to get away with it.
No need for fake ones since so many consider it their privilege to inform us . . .
So perhaps a paper is not the best way of testing or evaluating knowledge or thought. After all, after graduation how many papers will most people write?
Exactly this, IMO. 
lol, very true, and for most of these āinfluencersā its a true privilege to their wallets and pursesā¦
Good point. My push back would be these papers teach critical; thinking and formulation of complex arguments that can withstand critiques.
Very helpful if you are looking to publish in a rigorous academic journal, or write a book, or both.
I would also add the ability to find truth and true arguments, really help cut through the crap of all the advertisements, lies, misinformation, and self serving influencers out there.
I agree, writing is a form of thinking and reading is symbiotic with thinking, however most of what takes place in academia is more like driving bumper cars at the carnival than traversing an open road with a bike or an automobile.