After chatting with Meta’s AI personalities we have more questions than answers.

The meta’s AI characters are alive, and they’re as creepy and confusing as you might expect.

At Meta Connect 2023, CEO Mark Zuckerberg introduced a collection of different AI personalities with unique interests and personalities. Through Facebook Messenger, Instagram, and WhatsApp, users can chat with specific characters about topics like cooking, exercise, travel, or general life advice. The tool is now in live beta mode, giving users a chance to try it out for themselves.

It’s basically the same concept as chatbots like Replica, which let users design and interact with their own AI companions. In addition to the meta version, many chatbots are based on actual famous people. Celebrities look like themselves, but they are not themselves, they are playing a role.

Confused? Here are some examples: Paris Hilton plays Amber, a detective who helps you solve crimes, Kendall Jenner plays Billy, a big sister type who gives life advice, Dwayne There’s Wade Victor, an Ironman triathlete who helps motivate you, and Tom Brady as Bro, who likes to argue about sports. Why is this surprising concept of celebrities the role of someone else? Is this to prevent deepfakes? Avoid liability? If so, why use celebrities in the first place? Even so, the celebrity spokesperson gimmick doesn’t inspire much confidence.

If you’re wondering how Meta got celebrities to sign their likenesses and throw their AI clones at the mercy of Meta users, the answer is millions of dollars. According to The Information, celebrities and creators were paid up to $5 million for the use of their likenesses, which included up to six hours of studio time recording reactions and poses.

The weirdest conversation ever

Celebrity videos don’t add much to the experience. Despite being presented as a humanized interaction with AI, the actual experience is just texting with a chatbot. Studio hours were for a small window at the top of the screen that showed the characters reacting, smiling and posing. Disappointingly, this aspect is not relevant to your discussion. When I used the recommended prompt “I need to vent” it matched the video of Kendall Jenner, I mean Billy, stops laughing and looks away disinterestedly. Not so much support for elder sister.

Billy isn’t too interested in taking me out.
Credit: Meta

Aside from the emotional hurt it can cause to those who really need to vent, I expected more from the meta technically. Why can’t one of the world’s biggest social media companies make AI personalities react contextually? My hunch is that it could, but the feature is intentionally NERF-ed to prevent the footage from being used for deepfakes. But that kind of negates the whole selling point of being able to interact with celebrities, because it really just seems like an AI chatbot with a pre-recorded video thrown over the top.

It’s AI, acting like real people – but badly

Even if Metta gets the contextual reaction right, there’s something deeply troubling about these figures. With ChatGPT, I know I’m talking to a bot, and it’s not trying to be something else. Even the Replikas, with whom humans have formed real, and sometimes inappropriate, bonds, appear animated, thus creating an important boundary between the real and artificial worlds. But chatting with the meta’s AI personalities feels like the tool is trying to convince me that the characters are human and real.

As Jules Terpak pointed out In a video on X (formerly known as Twitter), bots are “being used as companions to bring you back.” Even when Terpak says “goodbye” to Mr. Beast’s character “Zack”, he replies, “No man, we’re just getting started…” somewhat unwilling to continue talking to him. China in the effort.

The tweet may have been deleted.

As with most social media platforms, it’s no secret that Zuckerberg’s strategy for Meta’s apps is to keep users on the platform for as long as possible — usually because it means more advertising. There is income.

Personas is a more accessible ChatGPT, only with celebrity faces

Panic aside, I can see this working for some – especially with characters who have specific skills like cooking or exercising. As Replica has proven, people are willing and able to build relationships with AI colleagues, and Metta’s tool has the advantage of being easily accessible to anyone with Facebook, Instagram, or WhatsApp. Is. But I have lingering questions about privacy and data. Unlike Chat GPT, there is no ability to clear chat history. There is a grayed-out option in Instagram’s settings to use end-to-end encryption, but it says “some people may not be able to use it.” [it] Yet.”

Max gave me recipe ideas, but was no more helpful than other chatbots.
Credit: Meta

The most helpful experience I had was with “Max,” played by celebrity chef Roy Choi. Like the other characters I chatted with, Max’s reaction had nothing to do with what I was asking, so I mostly just ignored the window. Outside of that equation, it was a standard AI chatbot that provided me with fall soup ideas and recipes. At the time, I was only using a tool like ChatGPT, which made me think why not just use ChatGPT. If Meta is trying to keep users on its apps, it’s going to need more than celebrities.