PPT Grading Changes (ChatGPT and AI Generated Submissions)

[zeagle1]

Just tosssed all my PT's from last season through ZeroGPT and everything came back at 0%. I think some of y'all are just self reporting.


What about AI generated graphics for PT's? Or do we already have something in place for that?

At least for graphics grading, a fully AI generated graphic is a 0/7. It can be murky when someone uses an AI-powered filter (to pixelate the sig for example) because clearly some work was done. I naively believe that most people will follow DBAD, but the league needs to remain vigilant for anyone who exploits this.

[shrub02]

Just checked the PT that I just posted against zeroGPT. 100% hand jammed. It came back as 48.36% AI generated. Apparently I write like AI.

[.Laser]

Agreed with all the points already made in this thread. It is perfectly reasonable for the league to police AI-written PTs in the sense of fairness and in the sense to prevent cheating. However doing so with a tool that has very clearly been proven to be shaky at best and is very obviously not perfect, flagging common phrases that many people use on a regular basis and picking and choosing what to say is and isn't AI-generated, is extremely unwise. As has been stated in the thread, this change IS GOING TO DRIVE USERS OUT OF THIS LEAGUE OVER A 150 WORD PT THAT AN AI BOT WRONGLY STATED WAS WRITTEN BY AI. The caps is to get the point across - we already have a very small userbase as the nature of the league and it is one that has retention problems already, as any hobby will. Using a system to screen AI-generated PTs is perfectly fine, but if we do it, it cannot be based solely off of an extremely flawed mechanism. Perhaps get another AI-detection bot and run it through both, then take the averages? Anything would be better than this. I VERY STRONGLY urge PPT and Head Office to reconsider this decision and come back at the beginning of next season with a different, more accurate way of doing this.


I wrote that personally and got these results. I did not use a AI bot to write any of it. As stated, this is an extremely flawed system and I very very strongly urge that this rule be reconsidered.

[MouseMazing]

Here is my English 401 nine page final essay (from 2020) run through zeroGPT, cutting it close! Also my first draft for this essay was "63.48% AI GPT*"