Watch your text! Big brother Blizz (and Microsoft) could be watching...

[Kilpi]

Weird.

Because I told a griefing pally in a pug to "go f*ck yourself" last night and there weren't links or anything. In all seriousness, I legitimately haven't seen anything like that, except that sometimes it censors what people say and provides an option to see what the text says. I haven't seen the report feature added to it, but I've also seen it so rarely that maybe I don't pay attention.

You didn't see a link to report yourself? Weird indeed.

[FossilFree]

It must have been way worse than OP is letting on, since they didn't feel courageous enough to share the text in question.

[deenman]

had a message like that poping for me,even if i have profanity disabled,it still blocked the message,and when i clicked to see it,it wasnt even a bad word,it also asked me if i wanna report it lol,my guess is blizzard started using AI for chat and its to dumb to tell what words are if writen wrong or may seem as potentialy part of something bad

[facelesssoul]

It must have been way worse than OP is letting on, since they didn't feel courageous enough to share the text in question.

Seems a lot of folks missed the actual quote in my OP, the person simply said "I hope you choke". The /w was not censored but it was immediately followed by a message from the game saying that it was "suspiciously inappropriate" with an in-line link to report.

What caught my attention was that the individual words in that message are not offensive individually. It could be argued that "you choke" could be a red flag but I could be saying "You choke on coke if you drink it lying down" and that could get caught. It seems like the automated system is now applying some level of phrase-level analysis as opposed to word level filtering.

The OP isn’t saying OMG CHAT ISNT PRIVATE!!11!

They’re saying it’s now automatically being *read*.

That is what I was trying to say in a nutshell.

[melzas]

Seems a lot of folks missed the actual quote in my OP, the person simply said "I hope you choke". The /w was not censored but it was immediately followed by a message from the game saying that it was "suspiciously inappropriate" with an in-line link to report.

What caught my attention was that the individual words in that message are not offensive individually. It could be argued that "you choke" could be a red flag but I could be saying "You choke on coke if you drink it lying down" and that could get caught. It seems like the automated system is now applying some level of phrase-level analysis as opposed to word level filtering.




That is what I was trying to say in a nutshell.

The "message from the game" was him whispering another thing, you can click on that filter to see what he actually wrote.

[Just Passing through]

There is always the same simple solution to every thread like this:

Stop being toxic to other people in game!

Op are you really that dumb?

This thing about not being toxic doesn't seem to be catching on, no matter the moderation tools....

[Relapses]

Just keep your proclivity for gamer words in your Discord. I have no idea why some people seem to view their ability to be a fuckwad to other people on the internet as an inalienable divine right. /shrug

[RobertMugabe]

So you’re ignoring that all WoW text is being fed into Microsoft’s AIs because it makes you feel superior?

The OP isn’t saying OMG CHAT ISNT PRIVATE!!11!

They’re saying it’s now automatically being *read*.

It's so funny, because you and OP just don't get it.

The chats were ALWAYS being automatically read, hence why the point of them never being private is so important. All messages have always been sent and stored in CLEAR text on Blizzard's database. We know that for a fact.

That's important to know, because this means that those messages may have been analyzed and flagged in the past already, probably since the very beginning (they may have improved over time). Just because you are getting some immediate feedback now doesn't mean they've only started doing this now.

[sam86]

way back when a website that used to exist dedicated for funny wow screenshots (i miss it, but so old i can't even remember its name) there was a screenshot of someone posting on forums claiming his account got hacked and blizz is doing nothing, a gm answered him (again on forums) that he was trying to buy illegal gold and sent his account details, the account owner faked ignorance, the GM flat out stated that they (as blizz) can see all chat even whispers
that info maybe over 15 years old now, while it was funny it showed that they at least record everything that we do or say in-game since pre-WoD era, do they do anything with it? probably not, just random text saved that probably get automatically deleted after 3 months pass or something, but they do record it
now AI is advanced enough to check all that random text to check what is offending here and flag it for action

[Doomcookie]

I'm not sure why anyone would be surprised that there's an automated system that tries to deal with hostile/abusive language. It's a cheaper solution than having hundreds of GMs taking tickets and digging through chat logs.

[Karlz0rz]

What caught my attention was that the individual words in that message are not offensive individually. It could be argued that "you choke" could be a red flag but I could be saying "You choke on coke if you drink it lying down" and that could get caught. It seems like the automated system is now applying some level of phrase-level analysis as opposed to word level filtering.




That is what I was trying to say in a nutshell.

A lot of words out there are fairly innocent, but once you string a full sentence together, they become vile :P.

Take the word "cheek" and "clap". Both very inncoent. Bring them together in a sentence, and suddenly you have something that people might take offense to ^^.
Automated systems will become more advanced as we move into the age of AI, so we just have to adapt.
Namely, behave a little bit better in game :P.

[Eurhetemec]

It used to be:
Something abusive said in chat---> Someone reports -----> It gets checked or a threshold of reports is reached ----> Action taken

Now it is:
ANYTHING said in ANY chat----> Client or Server checks for offensive content ----> Incident gets logged and escalated ----> You get prompted if you want to report ----> Action taken (not known if a report is even needed)

The bolded bit is a fiction you've made up, you have no proof or even hints that that is true, and it doesn't even make sense given Blizzard generally want to do less, not more.

So why make that up OP? Further, the "not known if a report is even needed" thing is hilarious paranoid nonsense. If they didn't need a report, bro, they wouldn't be putting the thing in some people can report it.

I get that you want to spread FUD, and make it like we should all be afraid mean ol' Blizz is looking over our shoulder, but the reality is, they aren't. They're far too lazy. The only people likely to be impacted by this are people who love to say truly nasty shit in chat.

You also made up a scenario in which another phrase "get caught" by this filter - but the reality is you could easily have tested that in-game, and seen if it happened, and we all know why you didn't - because you don't believe that phrase actually would "get caught" (re: drinking coke).

Maybe stop abusing people in-game if this really concerns you?

- - - Updated - - -

It's so funny, because you and OP just don't get it.

The chats were ALWAYS being automatically read, hence why the point of them never being private is so important. All messages have always been sent and stored in CLEAR text on Blizzard's database. We know that for a fact.

That's important to know, because this means that those messages may have been analyzed and flagged in the past already, probably since the very beginning (they may have improved over time). Just because you are getting some immediate feedback now doesn't mean they've only started doing this now.

Yup. That's probably how they developed this system, too - they've got two decades of training data!

Even back in the day, sometimes people would report someone and a GM would read the chatlog and find stuff in it that the person reporting hadn't considered might look bad for them.

Hell, I remember an incident from Cata or earlier where some guild were so dumb that they were discussing an exploit in-game, briefing people on how to use it, in guild chat, and were horrified (and tried to bring up similar stuff to the OP) when Blizzard had apparently seen this and they got in trouble for it (not just the people who had exploited, either - the people who were explaining how to exploit too). I doubt anyone reported them, and they doubted it too, they were like "Blizz is readin ur chat pls be afraid and feel bad 4 us" and I think the general reaction was "Well duh".

- - - Updated - - -

Just keep your proclivity for gamer words in your Discord. I have no idea why some people seem to view their ability to be a fuckwad to other people on the internet as an inalienable divine right. /shrug

Yeah it's truly bizarre. I've somehow played MMORPGs for 25 years now without telling anyone to kill themselves, die in a fire*, choke, or the like, and without calling them the n-word, f-slur or really any slurs or even insulting terms. I think the rudest I've got is using FFS in anger (which I think is milder in UK English than American English) or asking "What is wrong with you?!" to someone having a meltdown.

And I'm not even a particularly good person! I'd say I was kind of an abrupt speaker in chat, and I have serious ADHD!

So where other people get off on saying truly horrible shit, like really nasty stuff, I have no idea. And that they think it's their right is even wilder.

* = I'm not saying I haven't thought that at people. But I didn't type it! Or even say it to them on voice!