Are we at peak social media yet? It’s an interesting question to ponder after the launch of an iOS app offering a social media experience just for one (i.e. you). At a glance, SocialAI — which is billed as a pure “AI Social Network” — looks like Twitter but there’s one very big twist on […]
© 2024 TechCrunch. All rights reserved. For personal use only.
Are we at peak social media yet? It’s an interesting question to ponder after the launch of an iOS app offering a social media experience just for one (i.e. you). At a glance, SocialAI — which is billed as a pure “AI Social Network” — looks like Twitter but there’s one very big twist on traditionally microblogging: There are no other human users at all here.
SocialAI lets you share your thoughts with an infinite supply of ever-available AI-powered bots that can endlessly chat back.
Think about it: No remark you post to SocialAI will ever be greeted with silence nor fail to engage en masse. You simply can’t get ghosted. Because the app’s faux users exist to hang on your every word — leveraging programmed enthusiasm to chip canned commentary into your replies (even the sarcastic, snarky and pessimistic bots can’t resist joining these continuous scroll comment pile ons). And the best thing is you can be rest assured there are no actual humans to harsh your vibe.
We’ve seen human/AI social networking blends before — but this is a pure bot network (minus you). As the sole human in the chatroom, you are — by default — “the main character” in your own private online drama, each and every day you choose to log on.
While a bot-filled universe may sound like a fairly accurate description of Elon Musk’s X (formerly Twitter) these days, the big difference is that on SocialAI you can be 100% sure it’s 100% bots, 100% of the time. Because literally everything else is bot.
SocialAI is radically transparent that every user you encounter is code. It also pledges that all your faux social encounters are totally private. So while the interface apes the look of classic social media apps this is really something else entirely: A social space with zero prying eyes and a bunch of AIs to help you explore your curiosity.
Founder Michael Sayman, a New York-based, 28 year old developer who spent a chunk of his teens working for Facebook coding Snap-style features, calls the app’s vibe “liberating”. His company, Friendly Apps, raised a $3M seed round back in May 2022 before it had built any product. Discussing the response to the app’s launch late Monday he expresses surprise and delight that the MVP has triggered such an early buzz.
“It’s been crazy,” he tells TechCrunch. “It’s funny because I spent so much time thinking about like, okay, what should the invite flow look like? How should I, you know, try to help people — because I’d seen other social apps and they put such an emphasis on how to get people to share it with others… And then I just told myself I need to get this out the door.
“It’s a testament, I think, to the potential that it might have. So I’m just, I’m overwhelmed.”
Sayman says he’s been toying with the idea for SocialAI for years — wanting to build a safe space for people to share thoughts and get private feedback. But it was only more recently that AI technology, and specifically the large language models underpinning the boom in generative AI, caught up with his vision for a “private social network” where the user can bounce ideas off a diverse community of AIs.
“I saw these journaling apps and diary apps out there, but they all felt so empty,” he explains. “When people used them it just didn’t work any different than a notes app — and there was something about that that felt, you know, weird — and so, in that regard, I really thought there was potential in making something that just felt like a magical diary.”
SocialAI will feel instantly familiar to anyone who’s every used Twitter (or any of its clones). But scratch the surface and you’ll quickly see the facade of traditional social media drop away. Sure, just like any social app you can post comments, replies and like others’ comments but every interaction the app serves you has been artificially generated.
The most obvious point of difference vs traditional social apps is SocialAI users need to choose the categories of their “followers” — by picking from a long list of bot “types”. Such as “odd-balls”, “nerds” “intellectuals” and “trolls”, or “Liberals”, “Conservatives” and “Jokesters”, and many more besides.
You must select a minimum of three types of followers to populate your network — but there’s no upper bound; you can opt in to every single type offered (a full 32 at launch) if you want to really mix things up. (Or, well, as much as a feed of AI-generated content can truly range off pre-programmed rails.) Or you can opt to keep things cleaner and leaner with just a handful of types.
The types you pick will determine the flavor of the AI-generated chatter coming back at you. But Sayman also says the app is also designed to learn and adapt to its user over time, based on the sorts of followers and content you’re engaging with.
Want cheerleaders and lovers to cling to your thoughts? Select “supporters”, “fans”, “cheerleaders” and “charmers” and expect your banalest remarks to be overwhelmed with bottomless sycophancy. (“You look incredible!” “Oh darling, you look absolutely enchanting!” “Yasss, you look amazing!” etcetera, ad nauseam.)
For a little more spice with your artificial engagement, why not opt for a blend of “trolls”, “critics”, “sarcastic people” and “brutally honest” types? (“Wow, groundbreaking.” “Really? So basic!” “Please, don’t even get me started…”) Or to fish for more constructive advice, maybe “problem-solvers”, “ideators” and “teachers” will do the trick?
Just remember: You’re the only brain in control here. Everything else is programming. And if you don’t like the cut of certain bots’ jib just edit your types selection and try again.
The bots’ output absolutely tends towards full-blown cliché, as their types labels suggest. So the app experience can feel a bit like being waterboarded with motivational platitudes (i.e. if you’ve picked overly positive accounts); or a bottomless abyss of sealioning (if you’ve over-indexed on “contrarian” types). But, well, you do get what you ask for. And if you pick a mix of types you may be surprised to find some responses can stand out amid the more obviously artificial chatter.
You may also find yourself wondering how many hours of your life you’ve already spent arguing with bots on traditional social media.
On SocialAI, bot accounts are named true to type — hence you can chatter with “Rita Realist” if you want some home truths; or engage “Tina Troubleshooter” for tips on fixing stuff; or talk to “Connie Patriot” for the hot conservative take on of the day. (“How about a patriotic themed BBQ event to unite folks? Let’s celebrate American values and serve some good old hamburgers with apple pie!”) Nominative determinism sure makes a refreshing change from trying to spot bots on X/Twitter based on the quantity of numerals included in their account names.
Overall, despite the range of AI types on tap, SocialAI feels like a net affirmative experience. Even if you pick “trolls” and “haters”, these negative bots are — frankly — pretty polite by online standards. Definitely don’t expect vicious insults on demand. About the spiciest things got was when we managed to get one account to regurgitate a line at me that “fake people are the worst”. (Ho-ho.)
Sayman says the idea for SocialAI clicked for him when he thought about how generative AI interfaces are going to need to evolve in order for users to benefit from more diverse responses.
“When ChatGPT came out and I saw the chat interface, I thought to myself, well, this is cool, but surely we’re not going to stick to the chat interface for everything, right? And we’re here now, what, two years, three years later, every single AI app is just a chat interface. And I find it just really perplexing,” he says. “So I thought to myself, how do we help people interact with a lot of different AI models — because ideally I think people want to be able to compare answers. Especially in a world where these language models give you random answers every time you ask them a question. It’s not like they give you the same answer.”
“We don’t have to reinvent the wheel,” he adds. “We already have social media products that people use, that are somewhat intuitive for interacting with large numbers of people and users. So let’s build that. But let’s build the thing that ChatGPT should be.”
Sayman also claims he designed the app to help people “feel heard, to give them a space for reflection, support, and feedback” in a space that acts like “a close-knit community.”
The idea, he says, was inspired by his own need to have a sounding board of sorts when he was feeling isolated and had no one to turn to. “I know the app won’t solve problems for everyone, but I believe strongly that many people like me will use this to reflect [and] to grow,” Sayman explains.
In practice, the app feels more like a showcase for AI technology, and how it’s capable of mimicking how people speak and write, rather than something users would turn to regularly. (But it could at least be a safe place to post all those Twitter/X drafts you never had the courage to share!)
SocialAI is the third app to emerge from Sayman’s startup Friendly Apps, which has experimented with AI through an AI music streaming charts site, AI Hits, and in online dating, with Cosmic, which matches users with AI personality quizzes.
For now, SocialAI is a free download without in-app purchases. Sayman says he doesn’t intend to raise additional funds until he finds product-market fit.
Additional reporting: Sarah Perez
Leave a Reply