You are viewing a single comment's thread from:

RE: Forging AI Content

in #ai8 years ago

I agree that the content is easy to identify as nonsensical - at least to a native speaker. It may be more difficult to detect by someone who is translating. I'm also pretty sure that this type of content isn't being currently upvoted onto the trending pages either. However, it could possibly be used to generate a foundation for content, then be tweaked to make it more palatable to readers, particularly to lower-IQ readers or just skimmers. As this stuff improves, it would only make that easier to pass off as "original" and "acceptable."

Whether article spinners are being used or AI-generated content is, it's certainly possible to generate the feces, then shine it up for a feces post where 90% or more of the people voting for your post aren't even reading it. I'm not saying that it happens here, but it probably wouldn't be that hard to pull off.

(I think we should conduct an experiment?)

Sort:  

Polish the turd, in other words?

Yep. Some people are pretty good at polishing turds. If an AI program can do most of the heavy lifting, then it would only take a few minutes for a semi-literate person to make the sentences readable. I would argue that most people, especially when reading an unfamiliar topic, probably wouldn't notice logical errors and inconsistencies. Grammar, perhaps - but seeing the current state of education these days, I'm not so sure. Have you spoken to a millennial lately?

The thing is with this... it seems like each sentence has nothing to do with any of the surrounding sentences.

Look at the one on anarcho-capitalism... what is it saying? They're just a bunch of unrelated sentences that mean nothing.

Look at the one on anarcho-capitalism... what is it saying? They're just a bunch of unrelated sentences that mean nothing.

Some people would argue that anarcho-communism is the same way. Ha! Gotcha!

Coin Marketplace

STEEM 0.21
TRX 0.14
JST 0.030
BTC 69662.55
ETH 3388.96
USDT 1.00
SBD 2.77