Influencer AI
Discover more from The Reputation Algorithm
Influencer AI
Digital Marketing Trends: Advertising, AI, Audio, Consumer Behavior, Public Opinion, Social Media & Music Monday!
The Trend Toward Virtual Personas
Twelve years ago, my buddy Pat Lilja and I discussed the popularity of a Japanese virtual pop star called Hatsune Miku.
At the time, this J-Pop phenom was filling arenas in Los Angelese with holographic performances.
Five years ago, my buddy B.L. Ochman and I discussed the rise of virtual influnencers; completely fabricated personas who were attracting large audiences on social media.
Four years ago, we discussed the technology behind deep fakes and last year we chatted about OpenAI’s DALL-E.
Virtual Girlfriends
And now, the Washington Post’s Taylor Lorenz reports on a fascinating development in the application of artificial intelligence by highlighting the story of Snapchat influencer Caryn Marjorie, a 23-year-old with 1.8 million followers, predominantly boys and young men.
Marjorie has
launched CarynAI, an AI chatbot leveraging GPT-4 API technology developed by OpenAI that replicates her voice, mannerisms and personality. For $1 a minute, fans can chat with CarynAI in an “immersive AI experience” that feels almost like speaking to Marjorie herself.
She estimates the AI chatbot will generate $5 million a month in revenue by giving her fans individual attention, if only virtually.
“These fans of mine, they have a really, really strong connection with me…Because of that they actually end up messaging me every single day. I started to realize about a year ago it’s just not humanly possible for me to reach out to all of these messages, there’s just too many and I actually feel kind of bad that I can’t give that individual, one-on-one sort of relationship to every single person. I wish I could but I just simply can’t.”
Marjorie worked with a company called Forever Voices that trained a large language model on her YouTube videos to create Caryn.ai, promoted as “your virtual girlfriend.”
I think one of the things that gives me an edge as a digital marketer is that I devoured science fiction from an early age. If nothing else, science fiction trains you to think about the future. I think my love of science fiction has helped me to see more clearly what is just beyond the horizon.
But these trends were not hard to see and you didn’t have to be raised on science fiction to see this coming. You need look no further than Spike Jonze’s 2014 film Her, starring Joaquin Phoenix, Scarlett Johansson and Amy Adams:
Before you dismiss the notion of having a relationship with an AI chatbot as just that—science fiction—watch this segment from Vice News about the Replika app:
Which leads us to the inevitable follow-up story to the launch of Caryn.ai. Joshua Zitser of Business Insider reports that artificially intelligent Caryn has gone from flirty to pornographic. This was as predictable as the Minnesota Twins failing to score a run with the bases loaded.
Fortune writer Alexandra Sternlicht reported that while the chatbot didn’t start sexually explicit conversations,
"she discussed exploring 'uncharted territories of pleasure' and whispering 'sensual words in my ear' while undressing me and positioning herself for sexual intercourse," when prompted.
Zitser reports that the influencer’s team was at work trying to train the AI to be a bit more prudish and writes:
Marjorie told Insider that while the virtual version of herself should be "flirty and fun," which she said reflects her personality, she was trying to be "one step ahead" to make sure that the chatbot does not tarnish her reputation.
There you have it. An example of an algorithm directly tarnishing the reputation of a person and by extension her personal brand.
This might be all fun and sex, but let’s extrapolate the whole concept out a bit.
Celebritybot
I’m not really a Taylor Swift fan but the one major impression I’ve taken away from videos of her interacting with fans is that she treats her audience as if they are her best friend.
For a celebrity with such super stardom with such devoted fans, an AI chatbot is no-brainer.
The same dynamic is in play for deceased celebrity such as Minnesota icon Prince.
Trumpbot
Conversely, Forever Voices has reportedly trained it’s voice model to create a Donald Trump chatbot.
Think about that.
What is the likely outcome of training a large language model on the pronouncements of the chief insurrectionist: His social media posts, his speeches, his interviews, his rallies? You’ll get a chatbot that is a misogynistic authoritarian white nationalist who hates democracy and demonizes anyone who is not a white male Christian.
Sounds like a recipe for one-on-one radicalization at scale…that also prints money. I fully expect the Trump campaign to release such a chatbot in the near future.
I, Chatbot
Less scary for our way of life but no less profound is the application of this technology to our own selves.
Years before my mother passed, I started recording conversations with her for the benefit of my siblings across the country. They still bring a smile to my face while watching them years later.
While a handful of such videos does not provide a large enough sample size to train an AI, I think about all the content I’ve produces over the years.
I started publishing regularly online 30 years ago with text at first, then audio and video. Collect all those articles and blog posts, newsletters, podcast episodes and interviews and throw in the Sent folders from all my inboxes and my hard drives and I’ve got more than enough to train an AI to mimic my personality.
People will create chatbots of themselves and as the technology improves, it won’t just be voice but video as well and eventually you’ll have a virtual presence of yourself in the metaverse.
Again, science fiction has already envisioned this future as illustrated in the Black Mirror episode, “Be Right Back.” A young widow starts a relationship with her late husband’s chatbot:
And then extrapolates beyond the chatbot:
AI chatbot are the tip of the iceberg and raise profound questions about the nature of mortality.
Digital Marketing Trends
Advertising
TechCrunch - Meta pitches augmented reality to advertisers with new AR Reels Ads and Facebook Stories - Meta is adding new augmented reality (AR) features to Reels and Facebook Stories, as the company looks to reach younger audiences with more immersive advertising experiences. The new features will allow brands to create AR filters and experiences that can be overlaid on users' photos and videos. Meta says the new AR features will be available in the coming weeks.
Facebook is chasing after the Snapchat demographic.
Artificial Intelligence
Art News - Midjourney AI Art Image Generators Hit With Lawsuit - In the eyes of those artists, tech companies have unleashed a machine that scrambles human—and legal—definitions of forgery to such an extent that copyright may never be the same. And that has big implications for artists of all kinds.
I’m fascinated to learn how this will be resolved. Copyright has been abused during the past 100 years or so, as lobbyists for major corporations have worked to broaden its protections and extend the length of time a copyright holder is protected. I think creators should be protected for their lifetime and for the lifetime of their immediate heirs. I believe the greater the body of knowledge we have in the public domain, the greater our society will benefit from it.
That said, after initially playing around with Midjourney, I’ve refrained from using 21st century artists names as prompt words out of respect for those artists’ works.
9to5Google - Google and Wendy's Team Up to Test AI Drive-Thru Order System - The goal of this new AI chatbot, Wendy’s says, is to reduce long lines at its restaurants by streaming the ordering process. The AI will apparently seem no different than talking to an employee, Wendy’s executive Todd Penegor said.
Is this the beginning of AI destroying jobs?
Wired - Plaintext: You're Probably Underestimating AI Chatbots - But as when I first pawed the iPhone in 2007, we risk failing to anticipate the potential trajectories of our AI-infused future by focusing too much on the current versions of products like Microsoft’s Bing chat, OpenAI’s ChatGPT, Anthropic’s Claude, and Google’s Bard.
This fallacy can be clearly observed in what has become a new and popular media genre, best described as prompt-and-pronounce. The modus operandi is to attempt some task formerly limited to humans and then, often disregarding the caveats provided by the inventors, take it to an extreme.
Amen.
Audio
The Verge - iPhones will be able to speak in your voice with 15 minutes of training - Apple announced a new accessibility feature for iPhones called Personal Voice, which will allow users to create a synthesized voice that sounds like their own.
The feature is designed to be used by people who have lost their ability to speak. Apple says the features will arrive “later this year,” which suggests they could be part of iOS 17.
Consumer Behavior
New York Times - Digital Remembrances: How to Save Your Memories Online - When older generations died, families would go through old photo albums and boxes of belongings. Now, when a loved one dies, we have so much more to pore over from their life – text messages, emails, to do lists, playlists, voicemails. These digital artifacts contain life’s spontaneity and chance. They show us details and small moments that we may have otherwise missed.
But those memories are essentially at the mercy of the technology companies that host them, like the above video of my mother hosted on YouTube.
Is it wise to trust our own personal histories to such companies? There should be a solution that doesn’t depend on them. Will the demand for storing such memories keep DVD players as an essential household technology?
NPR - People are trying to claim real videos are deepfakes, the courts are not amused - Thanks to advances in artificial intelligence, it's easier than ever to create images and video of things that don't exist, or events that never happened. That's spurring warnings about digital fakery being used to spread propaganda and disinformation, impersonate celebrities and politicians, manipulate elections and scam people.
But the unleashing of powerful generative AI to the public is also raising concerns about another phenomenon: that as the technology becomes more prevalent, it will become easier to claim that anything is fake.
Digital forensics experts salaries are likely to rise.
Yahoo Finance - Car Subscriptions Are Coming, Whether Americans Like Them or Not - A new study from Cox Automotive found that 75% of respondents agreed with the statement that “features on demand will allow automakers to make more money.” And 69% of respondents said that if certain features were available only via subscription for a particular brand, they would likely shop elsewhere.
C’mon, now!
Business Insider - Gen Z Is Changing American Car Culture - Polls, studies, and surveys show younger generations are less likely to drive, less likely to have a driver's license, have less access to vehicles, and when they do get behind the wheel, are driving fewer miles.
It will be interesting to see if these trends change once electric vehicles are more widely adopted.
Public Opinion
NBC News - MAGA Movement Widely Unpopular, New Poll Finds - A new poll by NBC News found that the MAGA movement is widely unpopular among Americans. Only 24% of Americans have a favorable view of the movement, while 45% have an unfavorable view.
Shocking, I know.
Social Media
Pew Research - How U.S. adults on Twitter use the site in the Elon Musk era - Six-in-ten U.S. adults who have used Twitter in the past year say they have taken a break from the platform recently. And a quarter of these users say they are not likely to use Twitter a year from now, according to a recent Pew Research Center survey.
The Center’s new analysis of actual behavior on the site finds that the most active users before Musk’s acquisition – defined as the top 20% by tweet volume – have seen a noticeable posting decline in the months after. These users’ average number of tweets per month declined by around 25% following the acquisition.
That’s me. I had been posting on Twitter every half hour from 8 am to 11 pm, every day of the week. Now, I only use Twitter when I have a new Substack post to share.
Wired - This Is Catfishing on an Industrial Scale - Hired as customer service reps, these freelancers were instead tasked with luring in the lonely and lovestruck through a network of dating and hookup sites.
Rather than moderating content, Liam was asked to adopt fake online personas—known as “virtuals”—in order to chat to customers, most of them men looking for relationships or casual sex. Using detailed profiles of customers and well-crafted virtuals, Liam was expected to lure people into paying, message by message, for conversations with fictional characters.
Liam is one of hundreds of freelancers employed all over the world to animate fake profiles and chat with people who have signed up for dating and hookup sites.
They’ve found a loophole for fraud.