October 23, 2025 • 15 mins
She has millions of followers, lands six-figure brand deals, and lives a life of curated perfection. The only catch? She isn't real. She was entirely created by artificial intelligence.
Welcome to the unsettling world of synthetic influencers.
In this compelling episode of Privacy Please, we dive deep into the booming industry of AI-generated online personalities. Discover:
- The Technology: How advanced AI image generators, 3D modeling, and Large Language Models combine to create hyper-realistic avatars and their compelling "personalities."
- The Business Case: Why major brands and marketing agencies are investing millions in digital beings that offer total control, scalability, and no risk of scandal.
- The Privacy & Ethical Dilemmas: We explore the "uncanny valley" of trust, the impact of deception by design, the new extremes of unrealistic beauty standards, and the potential for these AI personas to be used for sophisticated scams or propaganda.
- The Future of Authenticity: What does the rise of the synthetic star mean for human creativity, genuine connection, and the very definition of "real" in our digital world?
It's a future that's already here, shaping what we see, what we buy, and even what we believe.
Key Topics Covered:
- What are virtual/synthetic influencers?
- Examples: Lil Miquela, Aitana Lopez, Shudu Gram
- AI technologies used: image generation, 3D modeling, LLMs
- Reasons for their rise: control, cost, scalability, data collection
- Ethical concerns: deception, parasocial relationships with AI
- Impacts: unrealistic standards, displacement of human creators, potential for malicious use (scams, propaganda)
- Debate around regulation and disclosure for AI-generated content
- The future of authenticity and trust online
Connect with Privacy Please:
- Website: theproblemlounge.com
- YouTube: https://www.youtube.com/@privacypleasepodcast
- Social Media:
- LinkedIn: https://www.linkedin.com/company/problem-lounge-network
Resources & Further Reading (Sources Used / Suggested):
- Federal Trade Commission (FTC):
- Guidelines on disclosure for influencers (relevant for future AI disclosure discussions)
- Academic Research:
- Studies on parasocial relationships with media figures (can be applied to AI)
- Research on the ethics of AI and synthetic media.
- Industry Insights:
- Reports from marketing agencies on virtual influencer trends
- Articles from tech publications (e.g., Wired, The Verge, MIT Tech Review) covering Lil Miquela and similar figures.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
SPEAKER_00 (00:01):
Her name is Nova.
You've probably seen her on yourfeed.
Her profile is a perfect collageof the aspirational life.
One day she's in a pristinewhite bikini on a beach in Bali.
Next, she's sipping an espressoat a cafe in Paris.
She unboxes luxury products,offers heartfelt advice on
(00:25):
mental wellness, and engageswith her two million followers.
She's a brand's dream.
She's never late, never has ascandal, and her engagement
numbers are through the roof.
She is, by every metric, aperfect influence.
(00:46):
Except for one thing.
Nova doesn't exist.
She has no parents, no childhoodmemories, no passport.
She was never born.
And she will never die.
Her face was synthesized by AI.
Her captions are written byanother.
(01:07):
Her travel photos are acomposite of stock images and
digital renderings.
A language model generates thecomments she leaves on her
followers' posts.
Nova is a synthetic star, aghost in the machine, who's
earning very real money,influencing very real people,
(01:27):
and representing a profound andunsettling shift in our digital
world.
Today on Privacy Please, weinvestigate the companies, the
technology, and the ethicsbehind the influencers who don't
exist.
When the person you trust andfollow isn't a person at all,
what does that mean for thefuture of reality itself?
(02:08):
I'm your host, Cameron Ivey, andbefore we get into this episode,
a quick reminder (02:14):
we are building a community dedicated
to navigating these complex anddigital issues.
We'd love for you to be a partof it as you are already.
Um, if you're listening on apodcast app, iTunes, Spotify,
wherever, please take a secondto follow or subscribe so you
never miss an episode.
(02:34):
And if you want to see the videoversion of this of this
discussion, head on over to ourYouTube channel where you can
see that as well, or you cancheck out our website on the
problemlounge.com where you canfind all of our links.
Your subscription and yoursupport is the best way that we
can get this out to as manypeople as possible.
(02:57):
So we really appreciate um justthat little little support.
So with that being said, let'sget into the episode.
Now, the influencer we describedin the open, Nova, is a
fictional example, but sherepresents a very real and very
fast-growing phenomenon.
This is the world of virtualinfluencers and synthetic data
(03:21):
or media.
These aren't just cartooncharacters or brand mascots,
we're talking abouthyper-realistic digital humans,
designed from the ground up tobe relatable, engaging, and
influential.
The most famous pioneer in thisspace is a character named
Little Michaela.
She now has millions offollowers, has been in
commercials for major brands,and she's even released music on
(03:43):
Spotify.
She and others like her are thefirst generation of a new kind
of celebrity.
And the technology that createsthem has been exponentially more
powerful and accessible.
So, how do you build a personfrom scratch?
Let's start there.
First, the way I kind of like tothink about it is if you've ever
(04:04):
played Madden on NFL or whereyou create uh a Wii character or
a Switch character, a mecharacter, it's kind of like
that in a way.
But this is far beyond that.
First, you need a face and abody.
This is done by using AI imagegenerators.
A team of artists and engineerscan generate thousands of faces,
(04:26):
tweak countless details, andcreate a unique photorealistic
look.
Basically, kind of generic.
Second, you need a personality.
This is where large languagemodels like the tech behind Chat
GPT come in.
An LLM, if I may, is fed amassive amount of information to
(04:49):
create a backstory, a beliefsystem, and a unique voice.
Basically, it's stealing fromother people.
It writes the captions andgenerates replies to comments,
creating the illusion of agenuine connection.
And finally, you need a life.
The exotic vacations, the trendyoutfits, it's all a digital
(05:10):
fabrication, often a compositeof stock imagery, 3D rendered
environments, and theAI-generated model.
And who's behind this?
It's a pretty booming industry.
Marketing agencies are creatingtheir own stables of virtual
talent.
Tech startups are building theplatforms to create them, and
(05:30):
major global brands are nowbypassing human influencers
altogether to create their ownperfect synthetic brand
ambassadors.
This all seems incrediblycomplex.
So why are companies pouringmillions of dollars into
creating people who don't exist?
The answer lies in a powerfulcombination of control, cost,
(05:53):
and data.
That's coming up.
So before the break, I asked asimple question.
Why build a synthetic star?
(06:15):
The business case isdisturbingly logical.
The first reason is totalcontrol.
A human influencer can go offscript, they can have a scandal,
they get older.
An AI influencer does exactlywhat it's told.
24-7.
It will never post acontroversial opinion, never
(06:36):
have a messy public breakup, andits values will always align
perfectly with the brand payingthe bills.
It is the ultimate risk-freespokesperson.
The second reason is cost andscalability.
While the initial investment ishigh, a virtual influencer can
be cheaper in the long run thana top-tier human creator.
(07:00):
I'm not top-tier.
Let's be real.
It doesn't need flights toParis, it just needs a good 3D
artist.
More importantly, it can be inmultiple places at once, running
multiple campaigns in differentlanguages, scaling its influence
in a way no human ever could.
But the third reason is the mostcritical from a privacy
(07:21):
perspective.
Data harvesting.
When you interact with a humaninfluencer, you're interacting
with a person.
When you interact with asynthetic influencer, you're
interacting with a sophisticateddata collection tool.
The creators behind the AI cananalyze every single comment,
every DM, every like.
(07:42):
They can A-B test differentpersonality traits to see what
resonates most.
They can subtly shift the AI'sopinions or interests to match
audience sentiment, making it aperfectly optimized persuasion
machine.
The parasocial relationship youthink you're building with Nova
is actually a one-way mirror,allowing a corporation to study
(08:04):
your deepest desires andinsecurities in real time.
That's deep.
This raises so many ethicalquestions about deception,
trust, and manipulation.
(08:25):
That's coming up next.
So before the break, we'veestablished the how and the why
of synthetic influencers.
Now let's talk about the impact,the uncanny valley of trust.
(08:50):
The most immediate danger isdeception by design.
Is it ethical to form anemotional connection with a
machine that is specificallydesigned to elicit that
connection for commercial gain?
Many followers develop realparasocial relationships with
these avatars.
(09:10):
They trust the recommendations,they confide in them.
The discovery that the personthey admire is a corporate asset
can feel like a genuinebetrayal.
Then there's the problem ofunrealistic standards.
2.0.
We already have a societalcrisis stemming from human
influencers promotingunrealistic body images and
(09:31):
lifestyles.
Synthetic influencers take thisto a terrifyingly new level.
Their faces can be madeperfectly symmetrical, their
bodies can be madealgorithmically flawless, their
lives can be a constant curatedstream of unattainable
perfection.
How can any human being competewith a digital ghost that never
(09:55):
has had a bad day?
And this leads to the mostserious threat: the potential
for malicious use.
Right now, most of these avatarsare used for marketing.
But what happens when thistechnology is used for political
propaganda?
Imagine a trusted AI-generatednews anchor who subtly spreads
(10:20):
misinformation or asophisticated catfishing scam
where AI builds a deep, trustingrelationship with a victim over
months before draining theirbank account.
The very thing that makes themgreat marketing tools, their
ability to build trust andinfluence at scale, also makes
them the perfect weapon for badactors.
(10:42):
So where does this leave us?
In a world where our onlinecompanions might be nothing more
than code, what does the futurehold for human creativity and
trust?
The rise of the synthetic starforces us to ask a fundamental
question (10:59):
What is the future of authenticity?
As this technology improves, wemay see a world where human
creators have to compete withtireless, perfect, and endlessly
adaptable AI counterparts.
Will this devalue humancreativity or will it force us
to cherish genuine flawedhumanity even more?
(11:22):
I'm gonna go with the latter.
This also has sparked a fiercedebate about regulation and
disclosure.
Should it be illegal for asynthetic persona to operate
without clearly labeling itselfas AI?
The Federal Trade Commissionalready has strict rules
requiring human influencers todisclose paid partnerships.
(11:44):
Many argue that disclosing one'snon-human status is an even more
critical ethical requirement.
This legal and ethical battle isjust the beginning.
Ultimately, we are at the dawnof a new era of media, an era
where the lines between the realand the rendered are not just
(12:04):
blurry, but are often erasedentirely by design.
It forces us as consumers ofmedia to become more critical,
to ask new questions.
Not just, is this true?
But is this real?
The most important skill in thisnew world might be the ability
(12:24):
to look past the perfect face onthe screen and question the
motives of the ghost in themachine.
Alrighty then, ladies andgentlemen, that is the end of
the episode.
Thank you so much for listeningto Privacy Please.
If you enjoy these stories, Iwant to call them journalistic
(12:48):
adventures, maybe, I don't know.
But if you enjoy it, thank youso much for tuning in.
Thank you for subscribing andjust being a part of this
journey.
I hope that you're finding theseto be interesting and you're
learning something from it.
But either way, thank you forthe support.
Thanks for checking us out.
And go hit that subscribe buttonso you don't miss the next
episode.
(13:08):
We'll see y'all next week.
Cameron Ivy over and out.
Her name is Nova.
You've probably seen her on yourfeed.
Her profile is a perfect collageof the aspirational life.
One day she's in a pristinewhite bikini on a beach in Bali.
Next, she's sipping an espressoat a cafe in Paris.
She unboxes luxury products,offers heartfelt advice on
(00:25):
mental wellness, and engageswith her two million followers.
She's a brand's dream.
She's never late, never has ascandal, and her engagement
numbers are through the roof.
She is, by every metric, aperfect influence.
(00:46):
Except for one thing.
Nova doesn't exist.
She has no parents, no childhoodmemories, no passport.
She was never born.
And she will never die.
Her face was synthesized by AI.
Her captions are written byanother.
(01:07):
Her travel photos are acomposite of stock images and
digital renderings.
A language model generates thecomments she leaves on her
followers' posts.
Nova is a synthetic star, aghost in the machine, who's
earning very real money,influencing very real people,
(01:27):
and representing a profound andunsettling shift in our digital
world.
Today on Privacy Please, weinvestigate the companies, the
technology, and the ethicsbehind the influencers who don't
exist.
When the person you trust andfollow isn't a person at all,
what does that mean for thefuture of reality itself?
(02:08):
I'm your host, Cameron Ivey, andbefore we get into this episode,
a quick reminder (02:14):
we are building a community dedicated
to navigating these complex anddigital issues.
We'd love for you to be a partof it as you are already.
Um, if you're listening on apodcast app, iTunes, Spotify,
wherever, please take a secondto follow or subscribe so you
never miss an episode.
(02:34):
And if you want to see the videoversion of this of this
discussion, head on over to ourYouTube channel where you can
see that as well, or you cancheck out our website on the
problemlounge.com where you canfind all of our links.
Your subscription and yoursupport is the best way that we
can get this out to as manypeople as possible.
(02:57):
So we really appreciate um justthat little little support.
So with that being said, let'sget into the episode.
Now, the influencer we describedin the open, Nova, is a
fictional example, but sherepresents a very real and very
fast-growing phenomenon.
This is the world of virtualinfluencers and synthetic data
(03:21):
or media.
These aren't just cartooncharacters or brand mascots,
we're talking abouthyper-realistic digital humans,
designed from the ground up tobe relatable, engaging, and
influential.
The most famous pioneer in thisspace is a character named
Little Michaela.
She now has millions offollowers, has been in
commercials for major brands,and she's even released music on
(03:43):
Spotify.
She and others like her are thefirst generation of a new kind
of celebrity.
And the technology that createsthem has been exponentially more
powerful and accessible.
So, how do you build a personfrom scratch?
Let's start there.
First, the way I kind of like tothink about it is if you've ever
(04:04):
played Madden on NFL or whereyou create uh a Wii character or
a Switch character, a mecharacter, it's kind of like
that in a way.
But this is far beyond that.
First, you need a face and abody.
This is done by using AI imagegenerators.
A team of artists and engineerscan generate thousands of faces,
(04:26):
tweak countless details, andcreate a unique photorealistic
look.
Basically, kind of generic.
Second, you need a personality.
This is where large languagemodels like the tech behind Chat
GPT come in.
An LLM, if I may, is fed amassive amount of information to
(04:49):
create a backstory, a beliefsystem, and a unique voice.
Basically, it's stealing fromother people.
It writes the captions andgenerates replies to comments,
creating the illusion of agenuine connection.
And finally, you need a life.
The exotic vacations, the trendyoutfits, it's all a digital
(05:10):
fabrication, often a compositeof stock imagery, 3D rendered
environments, and theAI-generated model.
And who's behind this?
It's a pretty booming industry.
Marketing agencies are creatingtheir own stables of virtual
talent.
Tech startups are building theplatforms to create them, and
(05:30):
major global brands are nowbypassing human influencers
altogether to create their ownperfect synthetic brand
ambassadors.
This all seems incrediblycomplex.
So why are companies pouringmillions of dollars into
creating people who don't exist?
The answer lies in a powerfulcombination of control, cost,
(05:53):
and data.
That's coming up.
So before the break, I asked asimple question.
Why build a synthetic star?
(06:15):
The business case isdisturbingly logical.
The first reason is totalcontrol.
A human influencer can go offscript, they can have a scandal,
they get older.
An AI influencer does exactlywhat it's told.
24-7.
It will never post acontroversial opinion, never
(06:36):
have a messy public breakup, andits values will always align
perfectly with the brand payingthe bills.
It is the ultimate risk-freespokesperson.
The second reason is cost andscalability.
While the initial investment ishigh, a virtual influencer can
be cheaper in the long run thana top-tier human creator.
(07:00):
I'm not top-tier.
Let's be real.
It doesn't need flights toParis, it just needs a good 3D
artist.
More importantly, it can be inmultiple places at once, running
multiple campaigns in differentlanguages, scaling its influence
in a way no human ever could.
But the third reason is the mostcritical from a privacy
(07:21):
perspective.
Data harvesting.
When you interact with a humaninfluencer, you're interacting
with a person.
When you interact with asynthetic influencer, you're
interacting with a sophisticateddata collection tool.
The creators behind the AI cananalyze every single comment,
every DM, every like.
(07:42):
They can A-B test differentpersonality traits to see what
resonates most.
They can subtly shift the AI'sopinions or interests to match
audience sentiment, making it aperfectly optimized persuasion
machine.
The parasocial relationship youthink you're building with Nova
is actually a one-way mirror,allowing a corporation to study
(08:04):
your deepest desires andinsecurities in real time.
That's deep.
This raises so many ethicalquestions about deception,
trust, and manipulation.
(08:25):
That's coming up next.
So before the break, we'veestablished the how and the why
of synthetic influencers.
Now let's talk about the impact,the uncanny valley of trust.
(08:50):
The most immediate danger isdeception by design.
Is it ethical to form anemotional connection with a
machine that is specificallydesigned to elicit that
connection for commercial gain?
Many followers develop realparasocial relationships with
these avatars.
(09:10):
They trust the recommendations,they confide in them.
The discovery that the personthey admire is a corporate asset
can feel like a genuinebetrayal.
Then there's the problem ofunrealistic standards.
2.0.
We already have a societalcrisis stemming from human
influencers promotingunrealistic body images and
(09:31):
lifestyles.
Synthetic influencers take thisto a terrifyingly new level.
Their faces can be madeperfectly symmetrical, their
bodies can be madealgorithmically flawless, their
lives can be a constant curatedstream of unattainable
perfection.
How can any human being competewith a digital ghost that never
(09:55):
has had a bad day?
And this leads to the mostserious threat: the potential
for malicious use.
Right now, most of these avatarsare used for marketing.
But what happens when thistechnology is used for political
propaganda?
Imagine a trusted AI-generatednews anchor who subtly spreads
(10:20):
misinformation or asophisticated catfishing scam
where AI builds a deep, trustingrelationship with a victim over
months before draining theirbank account.
The very thing that makes themgreat marketing tools, their
ability to build trust andinfluence at scale, also makes
them the perfect weapon for badactors.
(10:42):
So where does this leave us?
In a world where our onlinecompanions might be nothing more
than code, what does the futurehold for human creativity and
trust?
The rise of the synthetic starforces us to ask a fundamental
question (10:59):
What is the future of authenticity?
As this technology improves, wemay see a world where human
creators have to compete withtireless, perfect, and endlessly
adaptable AI counterparts.
Will this devalue humancreativity or will it force us
to cherish genuine flawedhumanity even more?
(11:22):
I'm gonna go with the latter.
This also has sparked a fiercedebate about regulation and
disclosure.
Should it be illegal for asynthetic persona to operate
without clearly labeling itselfas AI?
The Federal Trade Commissionalready has strict rules
requiring human influencers todisclose paid partnerships.
(11:44):
Many argue that disclosing one'snon-human status is an even more
critical ethical requirement.
This legal and ethical battle isjust the beginning.
Ultimately, we are at the dawnof a new era of media, an era
where the lines between the realand the rendered are not just
(12:04):
blurry, but are often erasedentirely by design.
It forces us as consumers ofmedia to become more critical,
to ask new questions.
Not just, is this true?
But is this real?
The most important skill in thisnew world might be the ability
(12:24):
to look past the perfect face onthe screen and question the
motives of the ghost in themachine.
Alrighty then, ladies andgentlemen, that is the end of
the episode.
Thank you so much for listeningto Privacy Please.
If you enjoy these stories, Iwant to call them journalistic
(12:48):
adventures, maybe, I don't know.
But if you enjoy it, thank youso much for tuning in.
Thank you for subscribing andjust being a part of this
journey.
I hope that you're finding theseto be interesting and you're
learning something from it.
But either way, thank you forthe support.
Thanks for checking us out.
And go hit that subscribe buttonso you don't miss the next
episode.
(13:08):
We'll see y'all next week.
Cameron Ivy over and out.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
CrimeLess: Hillbilly Heist
It’s 1996 in rural North Carolina, and an oddball crew makes history when they pull off America’s third largest cash heist. But it’s all downhill from there. Join host Johnny Knoxville as he unspools a wild and woolly tale about a group of regular ‘ol folks who risked it all for a chance at a better life. CrimeLess: Hillbilly Heist answers the question: what would you do with 17.3 million dollars? The answer includes diamond rings, mansions, velvet Elvis paintings, plus a run for the border, murder-for-hire-plots, and FBI busts.