The most overvalued technology in the history of the world has once again been revealed to be a cover for cheap human labor, the Scooby Gang at 404 Media reported today. In this case, Meta’s latest AI product, Muse, which purportedly makes calls for humans too afraid to call a barbershop, is actually just humans making calls to other humans. Meta’s chief AI officer Alexander Wang claimed last week that Muse was expanding to phone beta use for making outbound calls to US businesses. That AI could feasibly make restaurant reservations or, we guess, refill a ketamine prescription if it worked. But 404 has unmasked the ghost haunting Meta’s creepy, bot-ridden amusement park, revealing “a human agent layer” doing the heavy lifting. The social media giant, which hasn’t made a worthwhile product in two decades, would’ve gotten away with it too if it wasn’t for those snooping reporters.
An internal message board post to employees, reviewed by 404, shows the roundabout ways Meta obfuscates its human labor. “Muse doesn’t just dial a number. It calls a business on your behalf, handles the conversation, completes your request, and reports back with a transcript and a summary,” the post about Muse’s capabilities reads. “It also no longer works alone. Muse is now able to hand requests to a trained human agent, who places the call and works it through.”
Okay, so, Meta’s Muse AI makes calls on users’ behalf, communicates with the person on the other end of the line, and sends the user a transcript afterward, but only if a “trained human agent” makes the call and handles the communication. Does this simply mean Muse is an auto-dialer with transcription capabilities? It sure seems like it. But a much easier way to understand Meta’s latest product is to think of it like every other Meta product: Completely useless unless you’re a scammer.
Some Meta employees have been praising Muse on X, where AI boosters and AI agents alike can share ever more hype about the technology that’s boiling the Earth. Internally, though, Meta employees realize that a human-driven AI making calls for users has potential for “so much negative PR,” warning co-workers on the board that “It could portray us as ‘their AI is not good enough so they still need humans’ kind of coverage for this launch. Another post, with a fairly lucid handle of the situation, wrote, “Please PLEASE do not release this as-is and at least make this a user preference if we absolutely must launch this. Please.”
This is hardly the first time AI has been revealed to be nothing more than underpaid and hidden laborers, typically from countries like Venezuela and Kenya, where workers train and moderate AI models. Here in the U.S., that means people cleaning up human feces from the back of Waymos. Companies would like customers to think this is all magic, but as always, it’s simply exploitation that solves the problem of calling people.
Keep scrolling for more great stories.