Your new AI BFF

An AI Paris Hilton, Lil Miquela, and how gen AI has turned tech's age old problem into their new favorite feature

Welcome to Day to Data!

Did a friend send you this newsletter? Be sure to subscribe here to get a post every Sunday covering the technology behind products we use every day, rising trends in tech, and interesting applications of data science for folks with no tech background.

---

Last week, we talked CAPTCHA, which was one of the early ways engineers tried to tackle one of their least favorite things — bots.

Not all bots are bad. However, malicious bots “mimic human behavior and abuse business logic, allowing threat operators and fraudsters to perform a wide array of malicious activities” per Imperva. In 2022, it was estimated that 47% of all internet traffic came from bots (not all bad bots), a 5.1% increase from 2021. Bots are a scalable way to be a malicious actor on the internet, that yields more anonymity than a regular user on the web.

Everyone in tech has been fighting bots for decades. On social platforms, bots have been a nuisance. Elon Musk has been prolific in his anti-bot tirades, especially during his purchase of Twitter/X, where he declared “defeating the spam bots” was one of his top priorities at the company. His solution? “Authenticating all humans” and $1 subscriptions. In e-commerce, bots have been a nightmare for retail. “Grinch bots” are buying up inventory, especially around the holidays where 69.5M bot requests come across e-commerce sites daily, versus the non holiday average of 46.5M. In travel, bots are undercutting prices and stealing loyalty points. They’ve done a lot of damage, so it’s obvious we’d want them gone, right?

Generative AI has already had an effect on social media platforms when it comes to creating bots. As I’ve said, bots don’t have to be a bad thing. Social platforms have had non-malicious bots already. In 2016, LA-based company Brud, later acquired by Dapper Labs, launched Lil Miquela, which has been a whirlwind of watching confused Instagram users be baffled by the absurdism of her photos. Lil Miquela has been “photographed” with celebrities and posted sponsored content with brands like Chanel and Prada — all while being a personality created by digital artists and editors building hyper-realistic computer generated images.

Lil Miquela, a CGI social media influencer, edited to be photographed with the Olsen Twins.

Brud, Lil Miquela’s creator company, raised millions from Sequoia Capital and Founders Fund, both massively respected Silicon Valley venture funds. Lil Miquela has over 2.7M followers on Instagram. She even recently collaborated with BMW, as in their launch of the BMW iX2. More virtual influencers are popping on the scene, and brands are figuring out how to best utilize them in their marketing efforts. Oddly enough, consumers don’t seem to be as off put as I’d expect by virtual influencers. Consumers are following these characters, buying what they’re advertising, and eager to see what these influencers are doing in their completely fake lives.

In September, Meta released 27 AI personalities that were using the image of major celebrities like Tom Brady, Snoop Dogg, Charli D’Amelio, and Paris Hilton but infused with made up personalities to create AI personas that live on Instagram. For example, Tom Brady’s AI personality is called “Bru” who is a “wisecracking sports debater who pulls no punches”. Meta spun up an Instagram page for each of these personalities, where for now at least, most viewers are simply baffled in the comments asking “Are we SURE this isn’t Tom Brady??” Bru, who has the username gameonbru, has only 7K followers. It’s pretty confusing what the goal of these profiles are. During their launch, Mark Zuckerberg spoke about how folks can chat with these personas, message them for advice, and see what they’re up to in their everyday (but totally fake) lives.

Some initial questions from me include:

• If a Meta AI personality wants to do a brand deal, who’s negotiating the deal? Do the AI influencers have talent agents? Lil Miquela signed with CAA. Does the creator the persona is based on (aka Tom Brady) get a cut?

• Do these profiles count towards Meta’s daily active users? Are those numbers going to one day be broken down into authenticated human profiles and AI generated profiles?

• What did the deal look like for these celebrities to sign over their image to Meta? Estimates say one celeb got $5M. What if they want their image back? Would love to see the contract for one of these deals.

Meta’s lineup of AI personalities, all using the image of celebrities and social media stars.

Most people think this is Meta’s way of attracting younger users on the platform, as well as a way to retain the declining number of users on the platform. While I’m suspicious of the fact that much of anything will result from these AI personas, I think it’s an inflection point in the way big tech is viewing “fake” personas on their platforms. Years ago, launching 27 profiles of influencers that look exactly like real celebrities probably would have raised major red flags. Lil Miquela got away with it because she’s a fake persona all together. Meta’s profiles are confusing users more than they are attracting them. And if anything, they appear to incentivize the creation of the fake profiles that tech companies have worked hard to avoid. I’m eager to watch these platforms and their policies evolve over the next year in response to the rise of artificial characters and personas.

---

Thanks for reading! No bots were involved in the writing of this article.