Subject: AI Quiz Info
Posted by: lordprescott
Date: Jun 25 23
For the first time the other day, I saw that an adopted quiz bore a note that said that all interesting information had been added by AI. I'm not sure if this has been happening for a while now and I've just not noticed, and I don't want to start a debate, but I for one enjoy writing quizzes, and for ethical reasons I'd rather adopt as many as I can than have AI fixing them or doing anything else. Is this a new thing?
Posted by: lordprescott
Date: Jun 25 23
For the first time the other day, I saw that an adopted quiz bore a note that said that all interesting information had been added by AI. I'm not sure if this has been happening for a while now and I've just not noticed, and I don't want to start a debate, but I for one enjoy writing quizzes, and for ethical reasons I'd rather adopt as many as I can than have AI fixing them or doing anything else. Is this a new thing?
15 replies. On page 1 of 1 pages. 1
VBookWorm
|
I've also been seeing this and wanted to post about it, but LP beat me to it! I have a follow-up question though. Since these AI quizzes are still going up for adoption, does this mean that the experiment was a failure? Reply #1. Jun 25 23, 4:42 PM |
kyleisalive
|
The AI-generated info is the result of a test Terry's been performing on quizzes with zero info on the Adoption list. Should only apply to a select few. Those quizzes are still up for adoption; we encourage authors to adopt them nonetheless and improve them. The jury's still out on AI use and ethicality. The quizzes selected, because they had no info, were being penalized by search engines for having no content. They're from long-gone authors and the intent is to have them adopted either way. We're not encouraging authors to use AI to write quizzes for them. What we've been seeing is that AI, when asked, is generating some clear and obvious keywords that really come across as clunky, and with the technology constantly changing, it's not altogether ideal. It's not good at putting any sort of personality into a quiz and, worse, it's not good at extrapolating on clever ideas. For all we know, laws surrounding AI text could change in the coming years and it could land on something similar to plagiarism or, as noted, poor ethics, so we aren't set to say 'go ahead and ChatGPT your quizzes'. If it were to come to that, we would have to take all offenders offline, so it'd be a waste of everyone's time. In my opinion, there's not a world in which this type of technology doesn't continue to advance, but there will probably also be better ways to track and identify it. If and when it comes to that, it's my opinion and expectation that we would be incentivizing human authorship more. Reply #2. Jun 25 23, 5:07 PM |
| VBookWorm
|
Everything you said makes perfect sense, Kyle. Also, there are some things AI probably can't fix, like FITBs with punctuation. Reply #3. Jun 25 23, 5:23 PM |
agony
|
We won’t be getting rid of our human editors, or our human authors! AI appears to be not too bad at writing basic pedestrian prose. But that’s not really what we prize, here. No matter how this all shakes out, we will always prioritize creativity and originality. Not to mention wisdom and humour. Reply #4. Jun 25 23, 7:55 PM |
| VBookWorm
|
If the quizzes didn't have creativity and originality, much fewer thumbs would be given out! Reply #5. Jun 25 23, 7:56 PM |
lordprescott
|
Really true, VBookWorm! If however there does come a time when AI develops so that it does do a better job, I'd still opt for human authors any day. Reply #6. Jun 26 23, 6:34 PM |
| VBookWorm
|
I'm considering adopting an AI quiz but before I do, I wish for some clarification. Do I need to do anything special or just treat it like any other adoptee? Thanks! Reply #7. Jun 28 23, 5:32 PM |
| kyleisalive
|
To follow up on this, treat it like any other adoptee, but don't rest on the laurels of the contents being auto-generated. What we're finding is that a lot of these experimental quizzes contain quite a few errors resulting from the AI. Each one of these should be gone through with a finer-toothed comb because they don't have twenty years of plays to polish off those edges. Reply #8. Jul 10 23, 8:47 AM |
rossian
|
I discovered, thanks to a correction note from a very confused player, an artificially generated information section which said the given correct answer was wrong and gave a different one. AI can't pick up nuances in wordings (yet), so I'd certainly recommend checking every single fact if you adopt one of the quizzes with info created by computer. Of course, this advice applies to any other quiz too - I rewrote one last week which had at least two factual errors in it. It's sometimes less work to create a quiz from scratch than adopt. Reply #9. Jul 10 23, 10:01 AM |
| agony
|
It's starting to look like AI is a tool that could be useful especially for authors who are not completely comfortable with their writing and grammar skills, but it cannot be depended on as a "one and done" substitute for writing the quiz yourself. If you use it at all, use it as just one of the items in your toolbox, with both benefits and drawbacks. Reply #10. Jul 12 23, 9:23 AM |
LadyNym 
|
If I may add my own 2 cents, yesterday I played a quiz up for adoption that had AI-generated info because I was curious to see what it looked like. The best way to put the impression I had is that it felt soulless and repetitive, and that there was more filler than actual info. I ended up not adopting the quiz because it would have probably been more work to edit the II than to come up with something brand new. That being said, I agree with Agony that this AI-generated stuff could be a useful tool for those who are not completely confident about their command of the language. Reply #11. Jul 12 23, 11:05 AM |
| Terry
Head Honcho 
|
Yeah it was an interesting experiment but I stopped adding ai info after a brief experiment. It’s “ok” a lot of the time, but does come up with incorrect stuff occasionally. Probably in time they will improve these ai engines to truly be useful but for now they are definitely only marginally useful. Nothing is going to match real authors for a while to come at least! Reply #12. Aug 18 23, 11:55 PM |
FatherSteve
|
AI appears not to have developed a sense of humour. Reply #13. Aug 19 23, 11:45 AM |
| kyleisalive
|
Truly that. The problem with a lot of AI text these days it it lacks nuance. That might work for an educational regurgitation of facts for a book report, but it comes across as clunky and emotionless. Reply #14. Aug 19 23, 11:51 AM |
| lordprescott
|
I'm glad about that though. It means that so far it won't completely replace us! Reply #15. Aug 27 23, 5:22 PM |
15 replies. On page 1 of 1 pages. 1
Legal / Conditions of Use
|
|
: