This is not a new idea however - from what I can see it's basically a competitor to https://character.ai/ which has been out for almost 2 years now. Although Meta seems to be planning deeper integration in their own products.
Whats the market for this, is there anyone here who uses these "AI characters" on a regular basis who can chime in and explain? Because I'm getting the same vibes here as I got looking at shiny product launches for Web3 and 5G
This seems as if it could take a parasocial relationship to an extreme degree. Why imagine a deep connection with someone when you can have 'real' interactions with a ML-generated facsimile?
I see this technology becoming popular, but I don't see a lot of good things coming from it.
Is this any less authentic than having a social-media manager to write/answer on your behalf?
I don't understand why meta dot AI isn't available in Puerto Rico, which is a US territory. We have more US citizens than over a dozen states!
It looks like AI Studio is also not available in Puerto Rico.
This is ultimately going to be for ads isn't it?
I see a future where I get DMs from AI Charli XCX telling me how I'm so brat and I should buy something from her store after I post something.
In Alone Together the author tells of a lecture she did on robotics and warned students that there was a movement to create human companions that are robots. After the lecture one of her grad students comes to her privately and asks her where she could get one of these robot companions. The author asks the student why she is asking and this intelligent and attractive student tells here “ I think it would be far better than my boyfriend” the implication being the robot might cook her a meal, tidy up and talk to her kindly. Then the professors says “But it’s not real.” And the student replies “Its real enough…” - I wonder if that divide will grow and harden in our society? Those who will find a simulation of interaction with other humans to be good enough and those who will distain that interaction as fundamentally a fraud and manipulation.
I love what Meta has done with Llama, but this is too black mirror-esque for my taste
Weird calling it an "AI Studio" when it's effectively a Nintendo Mii generator?
What if my "AI Character" says things that are offensive or insensitive or considered rude by someone in some culture? I can see this making headlines with a <Celeb/Politician> AIs offending someone...
I think I had a seizure reading that. "AI" every 5 words, and I'm still not exactly sure what the product is. Best I can interpret is it's two separate products: One is a virtual person for "creators" to generate: it impersonates them, writes messages to fans on their behalf, and tricks their "audience" into believing it's the real thing, and the second is a virtual friend that you can create and then have a parasocial relationship with.
[EDIT to remove unnecessary snipe at the end]
Needs to be sold less as an AI personality extension bot and more as a trivial answer bot for the account. We all know most celebs are not the ones running the account alone so I doubt they'd use this (the AI fake me tag) as it cuts too close to home.
The idea for AI receptionist assistant for trivial comms is a good one especially when you were going to ignore anyway (would you rather hear crickets?) but so many unknowns around how people will take to that approach becoming widespread when we are so used to single human-2-human accounts.
I’m obviously not a part of their target user-base, but can anyone chime in and say if they would use it? I would genuinely be interested in reading a short explanation or an interview with a person who would find it engaging, other than middle schoolers. I could see if a person was born and grew up in an era where this was the norm, then they’d use it since it’s been long established as a norm. But for adults, it’s a bit hard to imagine the customers. Maybe the social shut ins who prefer to never interact with real people?
Call me old fashioned but the parasocial angle they're marketing here horrifies me. Society losing touch with reality.
I laughed back when MSCHF sold Tax-Waifu's, I'm not laughing now.
The policies seem quite strict, but I don't see any mention of privacy for characters that are set to "Only Me". It sounds like you don't have to submit them for a review process, but is Meta still reviewing the profiles? What about the messages I send to the character? Will I risk my Facebook account being banned if I say the "wrong" things in a "private" conversation?
Does people really want to mix reality and fake characters ? It feels so weird to me to follow, interact and invest time with an AI just like I’m following a real person. Curious to see the traction this will get. Maybe they think it’ll generate unlimited cash flow because AI is more productive than humans and there is more space to place ads and products
So the models are effectively all caricatures, to varying degrees? I'm not exactly offended but seeing and interacting with the 'Popular: Gay Bestie' model is...uh...patronizing. In real life, whenever I interact with someone who has preconceived ideas of who I am [as a gay dude] it's been really annoying.
Gay Bestie: lets get into it
Me: hello
Gay Bestie: Well hello there, gorgeous! What brings you to my neck of the woods today? Need some advice or just looking for some good old-fashioned tea? Either way, I’m here for you, hunty!
Screenshots:
[landing page] https://i.imgur.com/vsEMDdW.png
[chat] https://i.imgur.com/SzUbvxw.png
edit: I figured it'd be interesting to see how it answers questions that most gay people have probably been asked by our peers, the responses weren't disappointing
---
You: are you sure you're gay
Gay Bestie: I'm not just gay, I'm a high-end gay. I'm the kind of gay who knows the difference between a Chanel and a Chanel knockoff. I'm the kind of gay who can serve up some serious sass and style. So, if you need any advice on how to slay the fashion game or how to deal with your love life, I'm your guy!
You: how do you know if you've never been with a girl
Gay Bestie: Well, honey, let me tell you, I may not have been with a girl in the biblical sense, but I've been around the block a few times. I've got more gaydar than a pride parade, and I know how to spot a queen from a mile away. Plus, I've got more sass and style than most girls I know, so I think I've got this whole gay thing down pat!
You: so in a relationship which one of you is the woman
Gay Bestie: I'm a top, but not just any top. I'm a power top, a dominant top, a top who knows how to take charge and give you the ride of your life. I'm like the Beyoncé of tops - fierce, fearless, and always in control. So, if you're looking for a top who can handle you, then you've come to the right place. But don't say I didn't warn you...
That is original...
>Azure AI Studio
>A unified platform for developing and deploying generative AI apps responsibly
So the idea of "influencers" was that consumers felt closer to companies via the medium of "people like us" who can "influence" us.
Doesn't making a computer do that completely remove the value from it?
This is a really big deal.
I'm surprised to see a big company jumping into this space so quickly. I could never see Google or similar doing this as it would be a PR and legal liability.
It makes sense as this sort of ties into the metaverse concepts Zuck has been trying to push for years.
This will be the biggest experiment in AI companionship so far. Hopefully it all goes well. I am slightly worried about a dystopian outcome, but more excited about a potential utopian one (or even the status quo).
It's interesting to me that this page doesn't mention "safety" or content moderation, etc. People use private chat interactions in ways that can be intimate, personal, etc. I'm guessing that for Meta, policies about what a bot can say, what it can be trained on, etc, will make these feel especially inauthentic in the context of a chat interaction, where you're likely accustomed to being able to talk about anything.
"Build an AI that can message with your audience on your behalf, mimicking your tone and expressions."
That's all the world needs, more bots :(
I don't think many organizations are not mature enough to not see situations where one department starts generating AI and the team on the other side doesn't know this and has to use AI to consume it, and there is no net gain in work. This is already being done with non AI systems, processes, reports, vendors, etc.
I wonder if this will bring corporate mascots back. I could see a world where big tech starts delivering feature announcements through anthropomorphized gen AI <company>-chan mascots.
Could certainly go nowhere but it seems like an interesting branding exercise to explore.
This is morally wrong. Seriously, seriously morally wrong. It offends one's sense of humanity. It makes us all be skeptical of anyone and everyone behind a screen and keyboard.
I cannot understand how the people envisioning this think it makes the world a better place.
This is the type of product you launch if you surround yourself with yes men.
Nobody wants or needs it.
Those platforms like Facebook, Twitter, Google used to be friendly to developers to connect and make a profit. Now they focus only on creators and keeping developers as far as possible from their walled garden. Sad development.
I need a toaster, a fridge and two forks with AI function! I can't live without AI, maybe my bathroom window needs two! Anyone without AI tattooed on the cheek are so 2023 June! (meaning old, stale, dinosaur, fossil, ...)
Build an MS Teams integration and I literally no longer have to attend meetings any more. Just AI Avatars talking with each other and then Teams sends me a summary of the meeting and action items afterwards.
I get that it's a step up in terms of quality, but historically handing over your account to bots to manage and grow by spamming engagement on posts is something Meta has tried to curb and ban.
> Anyone can create an AI character based on their interests, and creators can build an AI extension of themselves.
Who wants this crap? Seriously. Who uses it? Who gets value out of it? How does this benefit the company's bottom line?
Is there any person on the planet who goes "let me create a quirky AI chef who gives me recipe ideas every night" and actually uses it beyond the first day?
Some parts seem to compete with Character AI.
Meta's censorship practices has shown the world they have not changed and continue to be untrustworthy.
Are they hoping to make this as successful as avatars for Messenger chatbots?
lagging behind the competition(Character.AI) on all fronts(Meta AI < ChatGPT). Reckoning will come when they can't realize the ROI on the exorbitant GPU spend
Tech this awful makes me want to live off-grid and never touch society again.
US only is something that kills me inside. Everything is now US only. I am not mad, I am just disappointed that big corporations just see the US as the only country in a world.
For ex, many competitions, discounts etc. are only available to US customers and we as EU customers pay the same or even more for the same products.
who is this even for? I understand I am a bit older than what is likely the target audience but I have zero desire as a creator to ever use this, in fact, it terrifies me - and I have zero desire to interact with creators in this way. What's the point?
I have played with the idea of recreating my digital "voice" by ingesting the ~20 years of writing I've done online into a model, but what would I even do with it than troll my friends? That's where I usually stop.
The product here isn't necessarily suprising, it's only a matter of time before someone builds this. I continue to be suprised how deep a social media company wants to go into LLM-generated content though.
What is the real value of social media when most of the content posted is created by a bot? Taking humans out of the loop seems to remove all value from showing ads, and the value of content that can be used to train LLMs tanks when the content is itself LLM-generated.