Youtube AI bans chess discussions

[mhlepore]

Kind of funny and kind of sad.

[baduk]

Kind of funny and kind of sad.

Well we need to give Ai some time,,I guess in 200 years it will be kinda accurate

[ez4u]

Interesting. Here is the Wired article. I don't like this repeated comment, "Fundamentally, language is still a very subtle thing." Does the professor think that YouTube and AI will manage to wring the subtlety out of language in the future? I think they should be saying the AI tools simply are subtle enough yet. Unfortunately, I expect that YouTube, Facebook, and so on will prefer to falsely ban some discussions on chess or go rather than let through more actual hate speech (for which they can be much more heavily criticized) by loosening their detection rules.

[baduk]

Im not sure if ai will be accurate ever,ai can not ubderstand things so, i cant see any improvement other than creating something beyond that

[Joaz Banbeck]

I'm a participant in a forum that has bots flagging posts, but the deletion - if any - is done by real people.
When a post is flagged, it gets put in 'hide' tags which function there much like they do here.
( For those who are not familiar with hide tags, please click on the one below )

This is text that can be hidden.

The viewer sees a note saying that the post has been flagged, and an icon to click on if they want to see it anyway. It remains like this until a mod looks at it - which is usually within 10 minutes during daytime.

I think that Youtube tried to take a shortcut. They had a system something like the above, tried to save on human labor, and gave too much power to the bot.
They don't really believe that bots can do the job, they just got lazy.