WEBVTT

[MUSIC PLAYING]

So a top venture capital firm is funding one company

to create a service that sends out thousands of fake influencers

to the interwebs, and simultaneously, that same VC

is funding another company to detect them.

So the marketing and content part of the internet,

it employs writers and video editors and musicians

and designers and other creative professionals like us.

And then brands, they hire editors and license

our music or biographics.

And all of this is created manually by humans using tools.

And many of us, myself included, have also

made this influencer content to market our own services

or our interests in related fields.

And even though we may have not felt comfortable,

it may have felt a little bit weird being called influencers,

like when our brand outreach was just

talking about dog training and making dog callers.

But now, generative AI services are

coming for a lot of this marketing work

that the creatives are doing, and it's

coming after the influencers while they're at it.

And we're seeing a dramatic increase

in these services in marketing that is fully generated by AI

from synthetic influencers all the way

to fake user-generated content or UGC.

And this rise is forming an emerging synthetic social web.

And in this episode, we're going to look

at how this synthetic social web is affecting

creative professionals beyond just inundating our platforms

with more stuff, with all that noise.

And the core issue here is the loss of authenticity,

in addition to the work, or perhaps whatever of it

is left on the internet.

And the whole point of UGC and influencer content

was that it was supposed to be a little bit more authentic,

a little bit more authentic marketing.

So what does that mean for the human internet

when a lot of this stuff is now becoming completely synthetic?

Hi, I'm Jen, and this is the Human Internet

Theory.

And in the show and podcast, I talk

about the changes to human creative content on the internet

and how it's affecting the creative professional industry,

what we need to know, and when possible,

what we can do in response to all of this happening.

So this synthetic avatar thing, it isn't entirely new.

For years, we've seen high effort or virtual influencers

like Lil Miquella.

Not sure if I'm saying that name right.

But you could call it the artisanal synthetic human.

But now we're seeing a move to just a really industrialized

synthetic human influencer.

It's a lot more accessible now because they're

going to be mass produced end to end,

the script right through to the editing process.

And this affects us creative professionals.

We are used regularly by the marketing industry

to strategize and design and write and plan

voice acting.

Now, one of these companies is a new startup

called DoubleSpeed AI, who is probably

about to help build this new synthetic social web

or kill what's left of the organic one.

This is a service that's backed by the prominent venture

capital firm, Andreessen Horowitz, or A16Z,

as part of their accelerator program.

Now, 404Media recently reported that DoubleSpeed

is selling thousands of synthetic influencers

to manipulate social media as a service.

Now, DoubleSpeed's stated mission on their website

is said to be automating attention

by building these synthetic creator infrastructures.

According to its company website,

these autonomous synthetic creators

have their personality set once,

and then they can go ahead and engage on social platforms

with full freedom.

So that means the posts, the comments, and the likes.

And the founder said on a podcast,

and I'll put the source in the description,

that they are using phone farms to operate

these AI-generated accounts on TikTok, for example.

Their business is said to be 95% driven by AI

with just 5% human refinement at the end.

So these services, as opposed to tools,

which tools you could potentially collaborate with

or use for creativity, these are services.

And this is important verbiage to get right in this AI era.

The generated output is a service.

So we're not talking about tools here.

Now, the tagline advertised on the website

is never pay a human again to address

and to automate what they see

as currently an inefficient process

or a human labor problem for brands.

This is right on the front page of their website.

And what's also interesting is that

while A16Z is funding Doublespeed,

a company that's creating fake online personas,

synthetic avatars for manipulating audiences,

the same VC firm also led a Series A investment

And Doppel's entire business model is using AI

to detect and remove the threats

that Doublespeed is creating,

like misleading impersonations and disinformation.

So perhaps the investors A16Z are hedging their bets

and wanting to sell both the disease and the cure perhaps.

I don't know, that might make sense.

So Doublespeed is generating influencers

for this new synthetic social media thing

that we're seeing emerge.

But let's look at another and very closely related AI service

type, which is a new group of companies

that want to replace UGC,

which is user generated content.

So UGC is a little bit different than influencers.

So influencers have their own audiences

that they've built up themselves around their persona.

And they're using that persona to promote

or market products using the audience

that they've built up around themselves.

UGC however, is typically content from users

or smaller creators that brands are given the right

to post themselves on their own brand accounts.

A UGC creator doesn't have an audience necessarily,

that audience isn't being purchased

or replaced in this case.

But very similar marketing outcomes,

it's real people using the product.

Now, creative professionals also work helping create

or edit UGC videos.

So this is another area where our work

is being replaced by services.

So these are companies that provide services for businesses

to generate entirely artificial content

that is explicitly designed to look and feel

like it was created informally by a real person,

often in their real homes or daily life.

So it's this unique style of video

that's now being replicated.

So let's look at a few of these startups.

So there's Creatify, AI, Zely AI and Arcads AI

who generate synthetic UGC videos

with really realistic AI generated avatars

to review the product in a UGC style.

Sometimes brands can just paste a product URL

and start this generation process from that alone.

Some startups are targeting marketing agencies directly

like CreatorKit, TagShop and MakeUGC.

MakeUGC directly says that they're looking

to replace the process of hiring human UGC creators.

And all of these startups have services like avatars

that will hold the product in hand

and are made high level to just eliminate the need

for photo shoots and filming and product users

and influencers and editors.

So they're eliminating the need

for creative professionals in the workflow

and they just say that on their websites

and I'll put that in the description as well.

So that's part of what's out there today.

And what's also out there today is my newsletter

and you can sign up for free at humaninternettheory.com.

I'm working on these segues.

Now, speaking of newsletters,

there are so many rules for businesses

and individuals around sending those out

and many more regulations around marketing, sponsorship

and transparency and those are in social media as well.

It's right across the board.

But now an entire industry segment is forming

around fake humans cosplaying real users,

vouching for paid product who are being made

by the people selling that product.

But regarding these synthetic human product users,

so-called, and this is just me,

but I do think it's time to change that term,

user-generated content for videos

that are synthetically generated by AI.

Gen AI, UGC just doesn't make sense as a term.

That term makes my brain melt a little.

It doesn't make any sense whatsoever.

Now, this tech was fairly patchy

and obvious as fake just a short time ago.

But now it's getting much closer

to being completely indistinguishable

from real people in a lot of cases.

And now this is a ridiculously small sample size,

I think of four, but an article I saw this morning

noted that these high school students

that were brought in for a news story

were only correctly identifying AI videos

as fake 67% of the time.

And all the while is turning into an actual segment

of the marketing industry that's centered

around just like breaking platform rules.

So we don't have legislation around videos online

that are copyright infringing derivatives

off of other people's videos.

But we do have all these big rules and penalties

around marketing emails, for example.

Because most of the platforms do have guidelines

against misleading people,

we gotta ask what we're doing here, right?

Now there are some news updates

around top down legislation changes in India, China,

and EU around labeling these videos,

which I'll cover in an upcoming episode later this week.

It's really important to keep marketing human centered,

I believe.

And for that, we need to keep humans

throughout the process and especially on screen.

The 95% AI, 5% human refinement

that's happening at double speed, as I mentioned earlier,

is pretty much just a tiny bit of QA at the end.

It has nothing to do with the creative

or the messaging part of the process.

And that does make me wonder what's going to happen

with like say liability down the road

in light of all of this,

not just what's in the terms of service,

but what actually ends up happening for real.

Now it's being really hard to watch the social web

go from actual friends and family, the social part of it,

even when it was just pictures of say dinner

and our dogs to largely something

that's about sales and marketing.

But now to watch even that part of it

possibly turn into a lot of synthetic

and kind of deception based marketing,

that just removes trust entirely.

And this could kind of kill the social web dead

because it hasn't been social for a while

or at least it might push it largely

into private communities instead.

And at the same time, it's removing a whole bunch

of our creative jobs and devaluing what work is left.

So the human labor problem that companies

like double speed are solving, I'm doing air quotes,

is the work of writers and editors and designers

and on camera talent and voice talent.

And the authenticity that UGC was all built on

is being just replaced by a cheaper synthetic alternative.

And that stuff is happening right now.

So I'll be back soon for the next episode.

And just FYI, I'll be focusing these videos

for a while on what's going on in the industry

and what tools we use.

I'm gonna keep things a little bit more

about the industry and it's so important

what's happening right now to the creative

professional industry and I wanna keep it real.

That's what I care about.

So maybe it's what you care about as well right now.

Bye for now.

- Episodes are written, directed, edited

and produced by Jen deHaan at StereoForest.com

Find out more about this podcast

and join our free newsletter for additional resources

at humaninternettheory.com.

Find additional videos at the YouTube channel

called Human Internet Theory.

Links are also in the show notes.

(gentle music)

(gentle music)