Meta is dreaming of its social platforms brimming with helpful AI-generated profiles and content fueled by the models underlying Meta AI, according to recent comments by Connor Hayes, vice-president of product for generative AI at Meta in an interview with the FT.
Of course, logging onto Facebook and seeing a “user” named Clara_ChefBot_9000 posting AI-generated pie recipes might sound like a great way to drive engagement to Facebook’s bosses, but it isn’t quite the same as scrolling through your Aunt Susan’s baking photos. Regardless, Meta’s vision of a platform brimming with relatable AI personalities is already beginning to roll out and if that doesn’t make you want to power down your devices and take a walk outside, let’s look closer at what this might actually entail.
The charm of social media has always been people. The whole point is seeing people sharing their lives. The oversharing, humble bragging, and dumb arguments about pineapple on pizza can irritate you, but at least you’re rolling your eyes about real people. AI-generated profiles, no matter how sophisticated, will always lack that authenticity. Sure, Clara_ChefBot_9000 might be able to generate recipes, but she’ll never know the joy of burning her first batch of cookies or the embarrassment of a recipe flop posted to a very judgmental audience.
Then there’s the issue of trust. We’re already grappling with deepfakes and AI-generated misinformation, not to mention people sharing a curated view of their lives as artificial as any algorithm. Do we really need bots with profile pictures trying to convince us of their “personal” opinions? Imagine arguing with a bot about politics, sports, or the aforementioned pizza topping, only to realize it’s been programmed to needle you into buying a product.
AI optimism
When AI assists humans in creative efforts, it can pull off some amazing feats, but when AI tries to mimic human creativity, it rapidly becomes dull slop or simply nonsense. Do we really need our newsfeeds clogged with bots sharing AI-generated memes or supposedly relatable status updates? And that’s if they don’t just go off the rails with errors. Imagine a feed full of AI profiles posting “Happy Mother’s Day” in the middle of November.
Meta AI can be fun (especially once you know how to turn it off), and there’s real value in an AI capable of entertaining you with conversation. Maybe Clara_ChefBot_9000 can offer you some cooking tips at midnight. That’s not a real, human connection, though. A chatbot might tell you how to bake a souffle, but it won’t commiserate with you when it collapses.
Meta might argue that these bots could help users with practical tasks, but that doesn’t mean there should be a million AI characters on the platforms. Social media platform SocialAI is attempting to sell people on that idea. Its mobile app connects each user only to AI chatbots they are supposed to interact with; they won’t encounter other humans there. That might be okay in limited contexts, but it makes no sense for Facebook or Instagram.
Social media exploded because it was a fantastic way to connect with people and make friends. AI can’t completely replace that. Instead of filling the platform with fake users, Meta should focus on enhancing the experience for the real ones. Otherwise, they risk creating a digital ghost town.