All hail the robot overlords. Automated censorship may be coming to IMVU. Together Labs is among a group of pilot customers testing and providing input on "Healthy Behavior AI". Together Labs is the current name of the company that owns IMVU, so this revelation may effect IMVU and its users. If you are not familiar with "Healthy Behavior AI", it is a content moderation solution that uses keyword detection across 30 different languages in order to automate the detection of "toxic behavior" and "remove bad actors". In other words, should IMVU complete testing and decide this is technology they wish to use, your private content and speech over IMVU may no longer be assumed safe within your controlled friends group and inner circle. No longer will the days of "snitches get stitches" and "oaths of silence" protect your freedom of speech. AI hast no honor among friends. What happens in Vegas no longer stays in Vegas when big brother is listening and tracking your every word. This new tech is said to track and analyze every users' behavior, keeping behavior profiles on users, and making AI decisions on moderation. A robot will be deciding if your speech and how you act is appropriate or not, programmed not just according to a complete stranger's view points on human morals, but on "brand protection". Spectrum Labs, the creator of "Healthy Behavior AI", freely admits on their website that their software is designed to "collapse content moderation costs by 50% and increase detection of brand-destroying behaviors". So the real question is, whose morals are these? - The morals of men, or the morals of the all-mighty dollar.

Sources:
https://www.businesswire.com/news/home/ ... ine-Gaming
https://www.spectrumlabsai.com/healthy-behaviors-ai

that's pretty disturbing... tho if vu implements this, they will quickly have to decide what's 'healthy'. is talking about sex 'healthy'? if the robot overloads are told that it's not, and that anyone talking about sex should get banned from the platform, how many users will they lose? how much VIP membership revenue goes bye bye? so then maybe the robots will be told that some sex chat is 'healthy' but not too much, or not certain fetishes, etc.

this raises another question, which is whether this kind of AI surveillance could be circumvented. as i understand it, chat messages on vu all go through their servers and get forwarded to the other users you're chatting with. if it gets implemented, the AI surveillance will happen on their servers. is there a way around this? like, the mafia or another org sets up their own server, and the clients are h4c|<7d to use that one instead? if the alternate server hosts content (avatars, rooms, etc.), the vu company could try to take legal action, but what if the alternate server only handled chat messages? vu can claim all the content is their intellectual property, but not our speech (well, they could try, but lol). some kind of hybrid solution might be possible.

another possibility would be to simply move to another platform. i looked at Meridian a little, but the avatars are really limited, and there's no store afaict. but Meridian shows that a better platform without the censorship/surveillance of vu could be done. there might be other platforms available now that are comparable, tho last i looked the only candidate i found was Meridian.

Ugh i've already seen so much moderation by these damn bots. Can't say anything anymore.. Our family Australian cant say "cunt" anymore without getting the "admin whispers" chat.. Got to love when snowflakes take over -.-

Damn' this is crazy. I bet this "Healthy Behavior AI" has not been trained on IMVU. So according to the AI, a lot of words revolving around sexual practices (which is, let's admit it, quite a huge part of IMVU content...) like "pussy", "cunt", "dick" or "ass" might be considered as insults. And would probably trigger the AI's moderation process.

On the short scale (just imagining which consequences this AI will have on our daily's experience on IMVU) : with a little investigation, I found out the Spectrum AI is already getting used for Riot Games or Fandom, which makes a huge field of testing... but completely unsuitable for IMVU. However, it seems they also use it for Grindr (a gay meeting app, which inspired Tinder), so hopefully maybe the AI already learnt, or could learn, the difference between saucy and rude words. At least we can hope.

But on the longer time scale, I'm as concerned as you because we cannot define ourselves what is morally acceptable, and as an outside tool will collect and analyze our datas in order to arbitrary decide it, we have no longer power on our datas, our conversation, and now our moral...

I wonder how it will react to the RP community. Does it understand the difference between IRL and fantasy?

Don Von Alpha Dom wrote:

@MickyBiggyI wonder how it will react to the RP community. Does it understand the difference between IRL and fantasy?

You're so right, that's also a good point of interrogation !

RP is (by essence) the ability to do impossible things in reality, and (like in Vegas) what happens in RP stays in RP. Personally, I'm not into extreme forms of RP (sadistic, humiliation, murder, gore, vore, r*pe...) ; but some people are ; and it's a good question if the AI will be able to make the difference between fantasy RP and insults/real threats. Most probably won't.

I could reply that an AI can be "trained", but considering how difficult it already is for humans to make the difference without any context, I highly doubt the AI would ever be able to make the difference. Good argument.

What's this bot lol is it kinda AI or something I've never heard about it frl Im kind anew to the game soo yea it's something that I know about well that's sound kinda interesting. So does the RP thing, like how does it work lmao.. I gotta explore more about this.