8

This answer at a glance looks like a generic new user's link-only answer. But if you look at the article it links to - signed with the same name as the answer - it seems to have been generated by AI. Someone got their hands on GPT-3 and had it generate the article.

At a glance it seems properly constructed, but on logical layer it's beyond nonsense - a soup of loosely related terms assembled into a grammatically correct structure resembling a tutorial article.

The answer itself seems to have been generated the same way, utilizing the first sentence from the article and a generic statement.

This is the first time I see something like this - bringing it to your attention, honestly I don't know what to make off it.

13
  • 1
    Doubt it's AI generated, but I will say that answers like this aren't exactly rare.
    – ave
    Oct 1, 2021 at 10:15
  • @ave: Could you skip down the article to somewhere near "This new method involves surrounding the device with a block of bedrock, " and explain how anything of this could be written by a sapient being? I tried adding an edit to my post explaining how an excerpt is wrong, but I gave up when I realized I'm addressing practically every single phrase with points of "why it's incorrect", "why even if it was incorrect, it would be pointless or undesirable", and "why it's not even on-topic" and "why nothing of it works anywhere close to that and/or is impossible!"
    – SF.
    Oct 1, 2021 at 10:48
  • 2
    I didn't click the link, I meant the post. It is 100% spam, I'm just saying that I saw spam answers like that before.
    – ave
    Oct 1, 2021 at 12:00
  • 1
    I am far more inclined to believe that the user simply copied and pasted their articles text and inserted the link and posted the answer manually. As ave said, I've seen similar posts in the past, across the SE sites. It just happens from time to time.
    – Timmy Jim Mod
    Oct 1, 2021 at 12:26
  • @TimmyJim The SE post itself might have been done that way, but the author of the post and the author of the article are the same name, and the article itself is definitely an AI work.
    – SF.
    Oct 1, 2021 at 12:32
  • @SF. About that, that it couldn't be written by a sapient (sentient?) being. I'm french. And even if I've been talking english for 20 years, I still sometimes explain stuff and it makes no sense for an english only reader. Just because sentences/points are built a little differently in different languages. Some twists in sentences or expressions do not translate well.
    – Fredy31 Mod
    Oct 1, 2021 at 14:07
  • 1
    @Fredy31 It's the opposite. The sentences are built correctly, and they are completely understandable, but the proposed ideas are a total nonsense. Surrounding the torch with bedrock block? It's unobtainable in survival, and completely pointless. Doors interrupting signal? (doors are strictly end-point devices, they don't influence the signal, they only can be influenced.)
    – SF.
    Oct 1, 2021 at 14:23
  • 1
    Or this: "The only difference is that if we place a single lever into an ocean of powered repeaters, it will act like a switch that breaks power to whatever device it's facing but does not require physical input at any time. " - 1. "difference" vs what? There's nothing analogous to compare to before. 2. The only way a lever can influence anything is through player right-clicking it so "does not require physical input" is nonsense. 3 It's a switchable signal source. On/off. It doesn't influence other signals, certainly doesn't break them. 4. "ocean of repeaters"??? What/why/uh... what?
    – SF.
    Oct 1, 2021 at 14:30
  • Well, it did happen to me a good couple of times to just mix up words and say nonsensical things. Some words sound alike or are literally the same in 2 languages, but don't mean the same thing. Example: Bureau. In english you even say it with kind of a french accent. Must mean an office right? No. A bureau is a desk. Anyways, I still stand with what I posted as an answer: Its bad quality and should be deleted, AI or not. And if an AI posts great questions/answers, I would then, at that point, sit down to clear that question of 'do we accept ai questions/answers'
    – Fredy31 Mod
    Oct 1, 2021 at 14:30
  • 1
    Now I think I know what astrophysicists feel when they read flat-earther papers, what lawyers feel when listening to SovCits, and what a microbiologist feels listening to a "natural therapy" anti-vaxxer.
    – SF.
    Oct 1, 2021 at 14:31
  • @Fredy31 A couple words maybe. But it's a lengthy article without a single piece of useful information - lots and lots of completely made up mechanics, analogies that aren't analogous to anything, and this all on a topic that can be entirely contained within a moderate-sized paragraph. Check my answer to that linked question, accepted. This is roughly the entirety of the subject - at least on the level the asker was asking.
    – SF.
    Oct 1, 2021 at 14:36
  • @SF. I'm now saying its a good blog. And really, at the end of the day, in the terms of Arqade, That its an AI or not doesn't do anything. Its a bad answer, and should thus be deleted. If one day, someone comes up with an AI that makes great answers, then we can have a discussion here. But right now, the only thing we need to do is see its a bad answer, and delete the bad answer.
    – Fredy31 Mod
    Oct 1, 2021 at 14:37
  • That article would be an amazing example of a parody of a tutorial article. It's very good unintentional comedy.
    – Fabian Röling
    Oct 9, 2021 at 15:12

1 Answer 1

Reset to default
3

This is a pretty common spam answer type. The 'Just let me plug my blog' answer.

And I do not think its an AI. The post is just a little bit in broken english, probably because the user's main language is not english (with the name, I would guess Indian)

We don't have a 'Does bruno mars is gay' situation here.

But that doesn't change the result, on our side. The answer is bad quality and should be downvoted/closed. AI or not doesn't change anything here.

Really wonder if we will have, at some point, to defend against AI Written Questions/Answers, but we will cross that bridge when we get to it.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .