Is it acceptable for zenimax to scour through your messages and take account actions at random?

spartaxoxo · October 5

the1andonlyskwex wrote: »

PrincessOfThieves wrote: »

JemadarofCaerSalis wrote: »

PrincessOfThieves wrote: »

Varana wrote: »

Most people here, even those who voted "No", know that things on the internet aren't really private, and that data storage and legal procedures for law enforcement are a thing. What this is about, is ZOS snooping around personal conversations and taking action even if there is no criminal investigation, harassment, hate speech, or anything that affects the public or legal issues.

This, pretty much.
If we use the aforementioned analogy with being invited to someone's house, getting banned for bad words in chat is more akin to renting an airbnb and getting kicked out in the middle of the night because you stubbed your toe on the way to bathroom and dropped an f-bomb.
Don't get me wrong, I am all for banning racists, sexists and those who act creepy and actively disrupt other people's gameplay. But banning people for just a few words is excessive and does not improve anyone's quality of life.
And what about other things that the AI might deem dangerous? For example, last time we had an endeavor that required you to kill X civilians, me and my guildies were discussing where to find them and how to do it without the police noticing. Should we all get banned for that, too? Should people who roleplay as evil characters get banned for mentioning things like slavery (assuming that other people in the group are comfortable with such topics)? In my opinion, this censorship thing is a slippery slope.

Thinking about it some more, I feel it might be more akin to buying a pass to use a club, and being given the club rules beforehand, and then stubbing your toe and dropping an F-bomb when it states on the rules that the club doesn't allow that type of language.

Ultimately, to me, it matters *what* language is being flagged and getting people banned for.

It may be because I have seen too many people claim they were banned for one thing, but in reality were banned for other things, or the people who sign up to a site that doesnt' allow multiple accounts/funnelling claim 'they banned both my accounts, and Ihaven't touched one in years! *then in tiny tiny letters buried deep in a sob story* except for once a month ago, because there was a pet that was really really sentimental and I wanted to transfer it to my new account!'

Basically, people lie about what they were doing that got them banned, mainly because they either don't want to admit they were actually doing something against the TOS or they want sympathy.

Well, we can compare the club with zone chat and the airbnb/hotel to whispers/group chat.
But the point is, even in your example the person will likely just get a verbal warning unless they were actively annoying others. A club is not a church service or something.
Clubs are businesses, after all, and they are not providing space and entertainment out of goodness of their hearts, and cutting ties after any minor offence is not a good business model.
Same with mmos, there is no real reason to ban people over super minor things when any account is a potential source of income.

Sometimes one of the selling points of a club is the fact that it's a safe space away from various unpleasant behaviors. In such a case, failing to police those behaviors (even minor instances) is what's bad business.

In the case of ESO, it very well may be better business to forego revenue from the most toxic members of the community (whose inappropriate conversations in "private" could easily be indicative of less obvious toxic behaviors in public) in order to create a more welcoming community for everyone else.

Sure, but how can you tell such a thing from the use of a bad word alone?

Syldras · October 5

Elsonso wrote: »

Well, apparently there is a new addon that obfuscates what a player says and only someone with the same addon, and a secret key, can interpret it.
I am sure that ZOS will catch onto that and find a way to ban anyone using it.

I can see them banning that - citing that they have to be able to check chats for criminal activity as a reason. And no matter how likely or unlikely people using that addon to discuss criminal plans is (plus there being many other ways not to trigger the filter bot just by phrasing things differently, using code words for obvious terms, or just using a language that's most probably not on the filter bot's list), it's hard to argue against that. Let alone they can ban whatever addons they wish on their own platform.

the1andonlyskwex wrote: »

Additionally, use of bad language can easily be a predictor of other bad behavior. For example, if you're calling your friends racial slurs as "jokes", you very well might also be treating suspected members of those races poorly in nonverbal ways when you encounter them in the game. ZOS could very easily have data that shows that people who speak in particular ways in private chat also have a tendency to grief opponents in PvP or to maliciously kick people from random dungeons.

Which brings us to the roleplay topic again: Just because kinlord Arildur calls another nobleman the imbecile son of a travelling wench (in his absence of course, because everything else would be indecorous) doesn't mean that the person playing him treats real players the same way.

Casul · October 5

Isn’t the entire point of the profanity filter to address this issue? If someone wants to avoid swears and slurs and see ***** as a replacement (I assume, admittedly I never use profanity filter) then why should AI monitoring be relevant. Leave the control to the players in my opinion.

With that being said this is only in reference to swearing and slurs. I have no comment on managing illegal activity on the in game chat.

the1andonlyskwex · October 5

Yes

Syldras wrote: »

Elsonso wrote: »

Well, apparently there is a new addon that obfuscates what a player says and only someone with the same addon, and a secret key, can interpret it.
I am sure that ZOS will catch onto that and find a way to ban anyone using it.

I can see them banning that - citing that they have to be able to check chats for criminal activity as a reason. And no matter how likely or unlikely people using that addon to discuss criminal plans is (plus there being many other ways not to trigger the filter bot just by phrasing things differently, using code words for obvious terms, or just using a language that's most probably not on the filter bot's list), it's hard to argue against that. Let alone they can ban whatever addons they wish on their own platform.

the1andonlyskwex wrote: »

Additionally, use of bad language can easily be a predictor of other bad behavior. For example, if you're calling your friends racial slurs as "jokes", you very well might also be treating suspected members of those races poorly in nonverbal ways when you encounter them in the game. ZOS could very easily have data that shows that people who speak in particular ways in private chat also have a tendency to grief opponents in PvP or to maliciously kick people from random dungeons.

Which brings us to the roleplay topic again: Just because kinlord Arildur calls another nobleman the imbecile son of a travelling wench (in his absence of course, because everything else would be indecorous) doesn't mean that the person playing him treats real players the same way.

Virtually all truly offensive real-world slurs aren't valid in-character RP slurs.

Syldras · October 5

the1andonlyskwex wrote: »

Virtually all truly offensive real-world slurs aren't valid in-character RP slurs.

Not the ones based on real-world races or nationalities, of course, but for example slurs relating to medical issues like mental or physical disabilities certainly would be (not that I use them, but I'm quite sure that an evil character might say really offensive things that would be absolutely inacceptible outside of roleplay). I don't know if you consider those "truly offensive", of course, I do.

the1andonlyskwex · October 5

Yes

Syldras wrote: »

the1andonlyskwex wrote: »

Virtually all truly offensive real-world slurs aren't valid in-character RP slurs.

Not the ones based on real-world races or nationalities, of course, but for example slurs relating to medical issues like mental or physical disabilities certainly would be (not that I use them, but I'm quite sure that an evil character might say really offensive things that would be absolutely inacceptible outside of roleplay). I don't know if you consider those "truly offensive", of course, I do.

I've never seen/heard any of those terms used by an NPC in-game either. They're not part of Nirn's vocabulary.

Syldras · October 5

the1andonlyskwex wrote: »

I've never seen/heard any of those terms used by an NPC in-game either. They're not part of Nirn's vocabulary.

The word from the last fable:
https://en.uesp.net/wiki/Lore:Blessed_Almalexia's_Fables_for_Morning
While that was a common term once, it would certainly be very offensive to call someone that today.

Amottica · October 6

Yes

spartaxoxo wrote: »

Amottica wrote: »

spartaxoxo wrote: »

Amottica wrote: »

spartaxoxo wrote: »

Groterdan wrote: »

Wauw the amount of yes . Rly surprised by that . Sad to see freedom gets given away so easy 😭

Again the problem with the poll is it's biased. I don't mean in a point of view sense. I mean that the selection of response chosen forces people to select responses that may not reflect their true opinion. This is why even most questions that are often a simple yes/no include an "IDK/other" option.

In this case, many users may not agree with your example ban or banning for things like jokes. But, they'd agree with narrowly defined monitoring for criminal activity. That's a pretty common response.

So you're seeing comments like "No, unless it's criminal" or "Yes, but only for criminal stuff."

If you had included a survey option that captured this common response type either by correctly predicting it or including an option for "other" then you may have received less "yes" answers.

We can choose yes or no; the question is a yes or no question with nothing to push a player toward either choice. That is unbiased by definition.

No. It isn't. Again, I'm not talking about a POV bias (leading question) that pushes players towards a particular choice.

"Forced-Choice" in polling is when a survey is created that limits the number of options that a person can choose. This can sometimes lead to different types of response bias such as acquiescence bias (people just agreeing even though the response does not accurately capture their opinion), especially when the forced choices does not accurately allow for respondents to convey an opinion that is more complex than the forced choices allowed. This is why it is often not a good idea to make a poll using only yes/no responses, it can lead to response bias in your data set.

We have that clearly going on in this thread. The OP gave the example he was thinking of in the comment section. Randomly being banned for banter with their friends, even though none of their friends reported the interaction or were offended by what was said.

Many of the "yes" responses indicate that they would ALSO be opposed to such random enforcement. They do however carve out a special exception for criminal conduct.

A more accurate poll would have either accounted for this likely response type and included it in the forced choices or included an "other/idk" option for respondents to elaborate further.

edit: further information

Image from "A Catalog of Biases in Questionnaires"

Source:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1323316/#:~:text=Forced choice (also known as insufficient category). Questions

Microsoft Copilot wrote:

Forced-choice bias. When respondents are only given a limited set of options (like “yes” or “no”) without the opportunity to express other opinions or nuances, it can skew the results and not accurately reflect the respondents’ true feelings or thoughts.

Not at all. It is simply a yes or no question that allows players to expand on their choice. The only thing that could have been added is "don't care" or "no opinion, " which does not make any aspect of the question biased. Not even close to it.

Not going to address this further as it is splitting hairs at best and not germane to the topic.

Respectfully, you are the one that decided to bring up the topic of definitions. So, I am merely clarifying that my word choice was accurate.
I have already shown textbook examples of how yes/no polling can lead to forced-choice bias resulting in inaccurate data. The OP expressed confusion about the amount of yes responses they received, so how yes/no polling can lead to forced-choice bias and acquiescence bias resulting in more yeses than would be expected is quite relevant. And it's not something everyone is taught. A lot of people make the mistake that a yes/no poll isn't biased, which is why it's a classic example for forced-choice bias. But it seems there is some misunderstanding about what the word bias means in polling. As you and I have both correctly noted from both of our very first responses, the OP didn't stick their personal POV into the poll. In polling, however, that is not the only definition of bias. I hope this provides further clarity.

When one suggests something is biased, they are the ones who started that conversation.

spartaxoxo wrote: »

I'm not going to vote because the poll is biased.

My answer is "sometimes" or "other" if those were included.

I understand scanning for certain criminal things to protect themselves from liability and also because it could save lives. I don't have a problem with that.

But, I don't think AI should be looking for things like rude language. If someone else reports you, that's a different story.

It was suggested twice before I replied to the suggestion.

I merely pointed things out to the contrary. For there to be bias, something needs to negatively affect the results. Nothing in the OP leads a player to choose one answer over the other, so there is no bias by definition.

SilverBride · October 6

No

I know they have a legal right to, but I find that degree of monitoring players to be way too invasive and unnecessary.

spartaxoxo · October 6

Amottica wrote: »

When one suggests something is biased, they are the ones who started that conversation.

Could have spoken about the results or the definition of the term, you chose the latter and the only one to do so.

I merely pointed things out to the contrary. For there to be bias, something needs to negatively affect the results.

The results were negatively impacted. As shown by some people answering yes or no but offering similar logic and stances.

Nothing in the OP leads a player to choose one answer over the other, so there is no bias by definition.

That is not correct. The lack of choices leads people to pick answers that do not necessarily reflect their opinion. At this point, I have to assume that you didn't actually read any of the things linked to you. You're ofc under no obligation to do so, but if you're not going to read what is linked than there is much point in discussing those links further. So I'll just bid you a good day.

spartaxoxo · October 6

SilverBride wrote: »

I know they have a legal right to, but I find that degree of monitoring players to be way too invasive and unnecessary.

Honestly, this is so true. Although, this degree of AI monitoring is so new that I don't think the legal framework is settled yet. But, even if they can do it, I don't think they should either.

Amottica · October 6

Yes

spartaxoxo wrote: »

Amottica wrote: »

When one suggests something is biased, they are the ones who started that conversation.

Could have spoken about the results or the definition of the term, you chose the latter and the only one to do so.

I merely pointed things out to the contrary. For there to be bias, something needs to negatively affect the results.

The results were negatively impacted. As shown by some people answering yes or no but offering similar logic and stances.

Nothing in the OP leads a player to choose one answer over the other, so there is no bias by definition.

That is not correct. The lack of categories results in people being led to pick an answer that does not accurately reflect their opinions.

Umm. Okay, I agree. I agree that not having a "no opinion" option could very slightly affect the outcome. However, any effect it may have is so minuscule that it is not something of concern.

Heck, it is the best poll question I have seen in these forums, and the lack of a "do not care" option does not seem to be a major concern.

As such, the poll results are worthy, though there is no reason for Zenimax to take notice.

Cheers.

spartaxoxo · October 6

Amottica wrote: »

spartaxoxo wrote: »

Amottica wrote: »

When one suggests something is biased, they are the ones who started that conversation.

Could have spoken about the results or the definition of the term, you chose the latter and the only one to do so.

I merely pointed things out to the contrary. For there to be bias, something needs to negatively affect the results.

The results were negatively impacted. As shown by some people answering yes or no but offering similar logic and stances.

Nothing in the OP leads a player to choose one answer over the other, so there is no bias by definition.

That is not correct. The lack of categories results in people being led to pick an answer that does not accurately reflect their opinions.

Umm. Okay, I agree. I agree that not having a "no opinion" option could very slightly affect the outcome. However, any effect it may have is so minuscule that it is not something of concern.

Heck, it is the best poll question I have seen in these forums, and the lack of a "do not care" option does not seem to be a major concern.

As such, the poll results are worthy, though there is no reason for Zenimax to take notice.

Cheers.

We can't know if it would have been slight because the poll was not conducted.

The most common opinions that I can personally see it a glance are

"They should monitor all chats for any TOS violations"

"They shouldn't monitor it because it's invasive etc."

"They should only monitor it for crime. They should not be checking for offensive language or other such TOS violations."

And that last one does not fit neatly into either yes or no. It could be viewed as "yes" because there is an allowance. It could also be viewed as "No" because the person doesn't believe that the vast majority of AI reporting is valid. And I've seen both yes and no votes express such an opinion in this or other threads. It is highly possible that this is the majority opinion rather than a blanket "yes" to AI monitoring.

That's why an "other" option was crucial to this poll since it does not seem that OP anticipated that as a response.

Rohamad_Ali · October 6

No one ever emails this one in game so not feeling eligible for voting here.

Amottica · October 6

Yes

spartaxoxo wrote: »

Amottica wrote: »

spartaxoxo wrote: »

Amottica wrote: »

When one suggests something is biased, they are the ones who started that conversation.

Could have spoken about the results or the definition of the term, you chose the latter and the only one to do so.

I merely pointed things out to the contrary. For there to be bias, something needs to negatively affect the results.

The results were negatively impacted. As shown by some people answering yes or no but offering similar logic and stances.

Nothing in the OP leads a player to choose one answer over the other, so there is no bias by definition.

That is not correct. The lack of categories results in people being led to pick an answer that does not accurately reflect their opinions.

Umm. Okay, I agree. I agree that not having a "no opinion" option could very slightly affect the outcome. However, any effect it may have is so minuscule that it is not something of concern.

Heck, it is the best poll question I have seen in these forums, and the lack of a "do not care" option does not seem to be a major concern.

As such, the poll results are worthy, though there is no reason for Zenimax to take notice.

Cheers.

We can't know if it would have been slight because the poll was not conducted.

The most common opinions that I can personally see it a glance are

"They should monitor all chats for any TOS violations"

"They shouldn't monitor it because it's invasive etc."

"They should only monitor it for crime. They should not be checking for offensive language or other such TOS violations."

And that last one does not fit neatly into either yes or no. It could be viewed as "yes" because there is an allowance. It could also be viewed as "No" because the person doesn't believe that the vast majority of AI reporting is valid. And I've seen both yes and no votes express such an opinion in this or other threads. It is highly possible that this is the majority opinion rather than a blanket "yes" to AI monitoring.

That's why an "other" option was crucial to this poll since it does not seem that OP anticipated that as a response.

Those answers are too convoluted, and some are a great choice to add strong bias. It is a simple yes or no question, and adding an "other" may have been nice so the current poll is a much better option.

spartaxoxo · October 6

Amottica wrote: »

spartaxoxo wrote: »

Amottica wrote: »

spartaxoxo wrote: »

Amottica wrote: »

When one suggests something is biased, they are the ones who started that conversation.

Could have spoken about the results or the definition of the term, you chose the latter and the only one to do so.

I merely pointed things out to the contrary. For there to be bias, something needs to negatively affect the results.

The results were negatively impacted. As shown by some people answering yes or no but offering similar logic and stances.

Nothing in the OP leads a player to choose one answer over the other, so there is no bias by definition.

That is not correct. The lack of categories results in people being led to pick an answer that does not accurately reflect their opinions.

Umm. Okay, I agree. I agree that not having a "no opinion" option could very slightly affect the outcome. However, any effect it may have is so minuscule that it is not something of concern.

Heck, it is the best poll question I have seen in these forums, and the lack of a "do not care" option does not seem to be a major concern.

As such, the poll results are worthy, though there is no reason for Zenimax to take notice.

Cheers.

We can't know if it would have been slight because the poll was not conducted.

The most common opinions that I can personally see it a glance are

"They should monitor all chats for any TOS violations"

"They shouldn't monitor it because it's invasive etc."

"They should only monitor it for crime. They should not be checking for offensive language or other such TOS violations."

And that last one does not fit neatly into either yes or no. It could be viewed as "yes" because there is an allowance. It could also be viewed as "No" because the person doesn't believe that the vast majority of AI reporting is valid. And I've seen both yes and no votes express such an opinion in this or other threads. It is highly possible that this is the majority opinion rather than a blanket "yes" to AI monitoring.

That's why an "other" option was crucial to this poll since it does not seem that OP anticipated that as a response.

Those answers are too convoluted, and some are a great choice to add strong bias. It is a simple yes or no question, and adding an "other" may have been nice so the current poll is a much better option.

I didn't say those should be the poll options, I said those are the three most common opinions expressed in this and other threads, and the last one doesn't fit neatly into "yes" or "no." Opinions tend to be more complex than a poll option should be

quinancia · October 6

Yes

We are still talking about the person who called his friend *** in chat and is upset about being banned for three days?

zaria · October 6

amig186 wrote: »

Amottica wrote: »

The_Meathead wrote: »

spartaxoxo wrote: »

I'm not going to vote because the poll is biased.

My answer is "sometimes" or "other" if those were included.

I understand scanning for certain criminal things to protect themselves from liability and also because it could save lives. I don't have a problem with that.

But, I don't think AI should be looking for things like rude language. If someone else reports you, that's a different story.

And honestly, this^

Someone messaging their pal with "Hey F-word, what's hanging?" is obviously a ridiculous thing to act on, but there probably are legitimate outliers.

There should be humans making the final decisions, and those humans should have a great deal of familiarity with context, familiarity, slang, and humor that they take into account.

Kevin stated that the AI only flags things to be looked at and takes no action on an account on it's own. Humans look into the flagged issue and decide how to act. He also stated they are learning from their experience and making changes.

I'm not sure if the humans doing the reviewing even get the proper context. My friend was suspended for 3 days for 'naming and shaming' when what he did was go on a bit of a rant, using words unsuited for polite company, about a particularly troublesome boss. In group chat, with only 1 other person whom he knew, so it wasn't possible that he was reported.

karthrag_inak wrote: »

Their system. Their responsibility. Their privilege.

I sure hope you never share anything personal in your 'private' chats with online friends, then.

I see it that they create an problem out of an non problem as in that getting said in private discussions.
But gaming industry has kind of grabbed the stupid ball hard lately.

quinancia · October 6

Yes

This is a long thread.

The owners of the game should define the terms of service.

The players should adhere to the terms of service.

If the players don't like the terms of service they should ask for change.

there should be no surprises.

katanagirl1 · October 6

quinancia wrote: »

This is a long thread.

The owners of the game should define the terms of service.

The players should adhere to the terms of service.

If the players don't like the terms of service they should ask for change.

there should be no surprises.

The surprises are when changes are made to the ToS. We have to accept a new one, but it’s a long document and there is no way to know what changed in it. There are no highlighted areas of change. It wouldn’t affect our ability to avoid them, but at least we would know what was coming.

Jaraal · October 6

Yes

"Don't do the crime if you can't do the time."

Sleepsin · October 6

Jaraal wrote: »

"Don't do the crime if you can't do the time."

Trash talk to friends is not a crime.

spartaxoxo · October 6

Jaraal wrote: »

"Don't do the crime if you can't do the time."

You've never used spicy language while joking with a friend? I don't mean slurs. But like saying a curse word because your RNG was bad or something?

Jaraal · October 6

Yes

spartaxoxo wrote: »

Jaraal wrote: »

"Don't do the crime if you can't do the time."

You've never used spicy language while joking with a friend? I don't mean slurs. But like saying a curse word because your RNG was bad or something?

What does what I do have to do with ZOS enforcing their TOS and Code of Conduct?

Jaraal · October 6

Yes

Sleepsin wrote: »

Jaraal wrote: »

"Don't do the crime if you can't do the time."

Trash talk to friends is not a crime.

Then why are people getting banned for it?

Make it make sense.

I mean, I couldn't care less what people say.... publicly or in private. But the OP is complaining about ZOS enforcing the terms of service they clicked "agree" to.

Sleepsin · October 6

Jaraal wrote: »

Sleepsin wrote: »

Jaraal wrote: »

"Don't do the crime if you can't do the time."

Trash talk to friends is not a crime.

Then why are people getting banned for it?

Make it make sense.

I mean, I couldn't care less what people say.... publicly or in private. But the OP is complaining about ZOS enforcing the terms of service they clicked "agree" to.

The op is complaining about getting banned when no living person reported him. He was talking to a friend in private channel and the AI took a key phrase out of context and flagged it for some intern to red and then ban him.

I_killed_Vivec · October 6

Yes

spartaxoxo wrote: »

Jaraal wrote: »

"Don't do the crime if you can't do the time."

You've never used spicy language while joking with a friend? I don't mean slurs. But like saying a curse word because your RNG was bad or something?

In context you can use the worst words in a friendly way... your friend has extreme good luck, adopt a cockney accent (think Bosmer for those unfamiliar with English accents) "You lucky c-word!"

Judgement is required (not to be used when talking to your mother-in-law).

spartaxoxo · October 6

Jaraal wrote: »

spartaxoxo wrote: »

Jaraal wrote: »

"Don't do the crime if you can't do the time."

You've never used spicy language while joking with a friend? I don't mean slurs. But like saying a curse word because your RNG was bad or something?

What does what I do have to do with ZOS enforcing their TOS and Code of Conduct?

Because that's what is happening with others. Like people are being banned for joking with friends in private chat. And it's not only stuff like slurs either. People got banned for telling off bosses or make risque jokes to friends.

BenTSG · October 6

No

If this is actually happening, then no, it absolutely should not be happening. Public chats such as Zone can be monitored for obvious reasons, but anything private such as DMs, or in a private space such as inside homes, shouldn't be monitored without a report when it's just friends talking to each other, or groups of people roleplaying.

ADarklore · October 6

No

Groterdan wrote: »

Wauw the amount of yes . Rly surprised by that . Sad to see freedom gets given away so easy 😭

This reminds me of Star Wars, "So this is how liberty dies... with thunderous applause".