You are welcome to look at the talkback but please consider that this article is over 2 months old before posting.
Forum - View topicINTEREST: X/Twitter's Grok AI Incorrectly Says Chris Beveridge Directs Laid-Back Camp Anime's Dub
|
Note: this is the discussion thread for this article |
| Author | Message | ||
|---|---|---|---|
|
RenimLS
Posts: 122 Location: North America |
|
||
|
Imagine all the power and resources dedicated to have the computers parse and process your question, dig up information, word your answer, and then to spit out a nonsense answer.
|
|||
|
|||
|
Sorraffy
Posts: 162 |
|
||
|
AI subtitles are going to be fun, hope crunchyroll and others are ready: release episode, multiple sub errors, fix, re-release. Next ep: release, errors, fix, re-release. Next ep: release, errors, fix, re-release.
|
|||
|
|||
Saeryen
 Posts: 939 |
|
||
|
I’m not surprised. AI tends to get things wrong about lesser known properties.
As for subtitles, I agree with what I believe is the majority opinion that AI should not be used for them. Machine translations don’t do well with Japanese often (especially with names) and I imagine that’s similar to how AI would translate. Plus the people who are already getting paid for this work shouldn’t lose their jobs! |
|||
|
|||
|
Lord Geo
Posts: 2589 Location: North Brunswick, New Jersey |
|
||
|
Well, should Crunchyroll give this dub a physical release in the future, one can only hope that Chris is at least given a "Special Thanks" credit.
|
|||
|
|||
|
PatrickD
Posts: 95 Location: California |
|
||
|
They need to pay Chris for his work!
|
|||
|
|||
Wyvern
 Posts: 1573 |
|
||
|
This is funny but also a good example of the potential AI has to unintentionally spread misinformation.
|
|||
|
|||
|
FilthyCasual
Posts: 2266 |
|
||
|
Okay, but what about the real news of basketball players throwing bricks through windows?
|
|||
|
|||
Raebo101
 Posts: 806 |
|
||
|
Honestly, I'd be kinda flattered if, say, I was mistakenly listed as Laid-Back Camp's English voice director. That dub slaps!
|
|||
|
|||
thekingsdinner
 Posts: 1081 Location: Geertruidenberg, Netherlands |
|
||
|
These sorts of issues are already becoming pretty regular, lol.
I laugh at this, but only because I'm not paying for any service at this moment and quit using Twitter, otherwise I'd just be sad. |
|||
|
|||
one3rd
 Posts: 1818 Location: アメリカ |
|
||
|
The fundamental problem is that these are language models and not actually Q&A bots.
They're designed to produce output that resembles natural language, not to give correct answers to factual questions. They will produce text that sounds like an answer to your question but they aren't capable of actually answering them. Whether it's correct or not is a coincidence that depends on the quality of the data the model pulls in. |
|||
|
|||
|
Twage
Posts: 361 Location: North Bergen, NJ |
|
||
Not sure about the "unintentional" part. X doesn't much care whether the information is accurate as long as it drives engagement with the site. |
|||
|
|||
 |
All times are GMT - 5 Hours |
|
|
|
Powered by phpBB © 2001, 2005 phpBB Group