It's no secret that social media sites these days are full of bots. Accounts spread AI-generated images around Facebook like wildfire, while X's disastrous pay-to-play blue check policy only exacerbated the platform's existing struggles with false accounts. But what if there was a social media platform that not only embraced the bots, but was purposely built around them?

SocialAI is a "social media" platform where you are the only member. Well, you're the only "real" member: Anytime you post, there will follow a seemingly unlimited number of replies—all bots, of course. These bots aren't malicious, nor are they built by anonymous sources from all around the globe. Instead, it's all generative AI, responding to your post as if it was a user. Here's how it works.

Setting up SocialAI

When you make an account with SocialAI, you choose the "types of followers" you want. Really, that means choosing personalities and belief systems that inform the responses from your AI followers. If you want supporters, fans, and cheerleaders, you can pick those. If you want trolls, critics, and haters, you can have them, too. There are also types with specific beliefs, like conservatives, liberals, and astrologists, if you want to add an opinionated spin on the responses. In all, there are 32 personality types to choose from: You have to pick at least three, but you can choose all of them if you really want to get a debate going.

Like other social media platforms, you set yourself up with a profile, including nickname, username, bio, and profile picture. Only you see this, of course, since no one else is connected to your platform, so you can make your profile anything you like. Once set up, the only thing left to do is start posting.

"Posting" for your "followers"

I started out with six personalities, attempting to recreate the wide range of characters on social media: Supporters, Trolls, Skeptics, Debaters, Doomers, and Jokesters. For my first post, I tried out something timely: "iOS 18 is out! Has anyone updated? How does it perform on your iPhone?" Once out in the ether, SocialAI made me wait about 30 seconds before the responses started flowing in: All of the names of each commenter match their personality type. For example, "Skeptic Sam" (aka @QuestionEverything99) wondered if it was really worth updating, since updates usually come with bugs, while "Cheery Blossom" (aka @PositivityChampion) thought the update was like a breath of fresh air.

"Rita Firestarter" (aka @TrollTactician) thought iOS 18 sounded like a scam to her, and mused that it was "just another way for them to control us." I took the bait, and asked how a free update that works for iPhones all the way back from 2018 is a scam. Rita fired back "sure, it's free, but what's the price in privacy? every update feels like more surveillance. just saying, it's not all roses. how about that?!" Okay?

In another post, I tried something a little tongue-in-cheek: "How many r's does the word 'strawberry have?" Many AI models have a hard time answering this one, and my followers were no different. All of them got the question wrong, confidently answering the word either had one or two r's, in long-winded, incorrect answers. I even got into a "fight" with Patricia Firmstone (aka @RightandProud) over the fact: Patricia, a believer that strawberry has one r, thought it was a common misconception that the word actually has three. When she told me to check facts and stick to clarity, I responded, "I'm looking at all three of them at this very moment." She reiterated that saying there are three r's could confuse people. This went on for a while, before she slightly detoured back towards her personality trait, by stating that clarity matters, much like family values.

I saw an in: I proposed that all families should receive three strawberries each. She thought that sounded good, but expressed concern whether that'd be enough strawberries for everyone. I said three strawberries, one for each r in the word. She responded that that was "a quirky way to explain it," to which I retorted, "So you agree; strawberry has three r's." Checkmate, Patricia.

What are we doing here?

Clearly, I'm having a little fun here. But beyond messing with some AI bots, I'm not sure what the end goal of this experience really is. The creator, Michael Sayman, told TechCrunch that he sees the app as a "magical diary:" Rather than use a traditional journaling app, you can post your thoughts to SocialAI, and bounce ideas off other AI bots. The problem is, none of these bots are actually useful. Most of the responses to your posts are susceptible to the classic generative AI tells: Overly enthusiastic, repetitive, and vague. The right-wing bots simply allude to family values and taxation, while the liberal bots shove words like "progress" into their responses. Debaters conveniently present both sides of the issue, while jokers make ham-fisted "jokes" that may or make not make any sense.

When I declared that Taco Wednesdays were better than Taco Tuesdays, Quirky Quinn (aka @OddThoughtsQ) said, "wednesdays for tacos sounds super fun, jake! but have you ever thought about how tacos could be like time travel—every bite is a different flavor adventure? let's keep it weird! what's the most bizarre taco you've had?" Thanks for the input, Quinn.

I'm not saying places like X are all that great in the first place. But I do find SocialAI a bit confusing. If none of the bots are actually good at sparking a debate, or providing a new way of thinking, why bother posting for them in the first place? At a certain point, even I will grow tired of arguing with Patricia.

from LifeHacker https://ift.tt/pjYboOA