12:04 PM
A: Is the Stack Exchange community going to fight against AI use with anti-cheating systems?

cocomacTo simply answer the question in the title (and what I think your overall question is): There are tools that exist to detect AI-generated content, although the specifics/details aren't public. They, of course, can have some limitations (although to make it more difficult to evade the ChatGPT/AI-c...

FWIW, Scott Aaronson has been working for OpenAI, developing techniques to robustly watermark ChatGPT output, but they haven't been implemented yet. See My AI Safety Lecture for UT Effective Altruism.
@PM 2Ring Problem is there will be a lot of AI systems open to public in the next years, not only ChatGPT.
@Universal_learner Sure, and we need to judge each of those systems on its merits. We can't possibly frame a general AI policy about systems that don't exist yet. OTOH, we can have a general policy towards pure LLM systems, and it's likely that the makers of all such systems would desire a watermark for their system, if it's possible. A watermark not only makes it hard to engage in plagiarism using AI output, it also makes it easier for the AI maker to avoid using the AI's own output in future training data.
@PM 2Ring Problem is there could be AI private and closed code products in the future not interested in anti-plagiarism. In chess they figth, but you will no see Magnus Carlsen playing 30' games in lichess. He plays 30'' or 1' not to be exposed to cheating.
@Universal_learner Why would a private AI company not be interested if a person uses the AI's output and makes money from that output by pretending that they created it themself?
12:04 PM
@PM 2Ring The true battle is going to be in the schools. And I guess if teachers have tools to detect all plagiarisms there will be products anti anti-cheating teachers tools in the market. I asked for example ChatGPT Can you write me a scholar work that sound written by a child? And it can. Scientific AI's may incorporate in the AI the anti-plagiarism tool. I asked ChatGPT to detect if a text was writeen by it and it fails to detec itself.
@PM 2Ring The student can even say the AI: Analyze my texts and write me a work about this topic with my own writing style.
@Universal ChatGPT isn't designed to detect its own output. Perhaps you should learn more about how it works, what it can do, and why it has the limitations that it does. There's some good info on Scott Aaronson's blog, and Stephen Wolfram's article is great, although it's aimed at people with some knowledge of coding & maths. writings.stephenwolfram.com/2023/02/…
@PM 2Ring No I have no math knowledges and I couldn't understand that. I am just a newbie that realized yesterday using ChatGPT for fist time that soon, when Google incorporates an AI in Android Studio, I will disclose my geology apps very fastly with AI code free of mistakes and I will not have more questions in this site.
I wonder how is this going to change the site but got not good acceptance
as fore me stackoverflow users -=1
@Universal_learner Ok, but I still recommend trying to read through Wolfram's article, and just sip over the parts with too much maths. It will help you understand why ChatGPT is so powerful, but it will also help you to understand that we need something in addition to a Large Language Model to achieve true AI that we can trust
* sip -> skip
Here's a crude analogy. ChatGPT is like a person doing a jigsaw puzzle. You give it a prompt of a few joined pieces and it adds more pieces, one by one. The picture on the jigsaw consists of a bunch of words written in English, and the jigsaw pieces are cut so that when you join them together the result will be syntactically correct English.
However, the guy doing the puzzle doesn't understand English. When we see the result, it looks impressive because we understand English, but ChatGPT just knows how to stick bits of jigsaw together.
12:36 PM
Interesting. I will read the article
As for me I only ask silly questions. I am part of the cleaning of garbage AI is gonna do in the site
I don't know why so much complains. To clean up the site and dedicate to top level programming challenges is a good thing
For me is like "Ok, I just wanted to publish my geology things, I din't want a degree in programming"
And AI is gonna help a lot, hoppe incorporated in ANdroid Studio soon
But it looks like some people is angry or has fear because they are not going to be reputation points available for the knowledges tehy have, only for the experts. That's afterall what the site is, a site for experts, not bad questions as mines
In Stack Exchange it is a different history with fun questions as "Will the AI formulate a unifying theory of physics?"
1 hour later…
1:43 PM
@PM 2Ring Thanks for talking with me, I have a menthal illness and it is sometimes hard. I am happy I will soon not need to ask anything more about my code. It is better Google does programm the code and nobody wastes his time lol :)
I learned how to programm at the end of my geology degree and I have suffered to publish my content in the store :)'
I don't know if that means I am ended as a programmer or that I can opt to apps development jobs as someone that knows how to programm manually!
something of the past
Once Google launches AI in Android Studio it will be time to publish. The store will fill of russian apps developped in two days. AS for myself I have the content: geology and I could profit that to publish lots of apps. Waiting patiently.
As for stack they will incorporate the AI answer as a warning before new users try to post
For my apps I associate with geology University teachers to do the content. And I could say a teacher "I can publish what you wish in two days"
a true advance
But it is true he can answer "I can do it myself lol :)"
So I am ended :)
Since 2015 doing courses for anything. I will be just another russian earning a bit thousands at the beginning but prograssivelly dying when designers start to open developper accounts
2:15 PM
@PM 2Ring you could find of interest, and maybe answer this question I come to post. ai.stackexchange.com/questions/39640/…
The big wonder for programmers should be if this technology can contribute to programming theory
if so, this is going to change a lot
if not, this is going to change a bit