Artwork for podcast Human Internet Theory
Building Trust as a Creator While AI Avatars Are Emerging
Episode 4 • 24th September 2025 • Human Internet Theory • Jen deHaan
00:00:00 00:28:04

Share Episode

Shownotes

The internet is quickly filling up with AI-generated content, including marketing that uses realistic AI avatars. Avatars, digital humans, can look and sound realistic and not have bad hair like me. In this episode, I explore the complex psychology of trust in the age of AI which gets into some of the hidden costs involved in using avatars and generated content.

I end the episode with three practical ways you can adjust your content in this new era by leaning into your humanity to build trust.

These episodes are recorded as a video. You can find the videos here on YouTube: https://youtube.com/@humaninternettheory

Resources Mentioned:

===

==========================

About and Support

==========================

Written, edited, and hosted by Jen deHaan.

Find this show on YouTube at https://youtube.com/@humaninternettheory

Subscribe to this show's newsletter for additional resources and a free 3 page workbook when you join https://humaninternettheory.com

Produced by Jen deHaan of StereoForest https://stereoforest.com

Contact Jen at https://jendehaan.com

==========================

Connect on Socials

==========================

Support

Your support will help this show continue. Funds will go towards hosting and music licensing for this show and others on StereoForest. This show is produced by an independent HUMAN artist directly affected by the state of the industry. StereoForest does not have any funding or additional support.

If you find value in our shows, please consider supporting them with a one time donation at https://stereoforest.com/tip

We love our podcast host Capitvate.fm! Contact me anytime to ask me anything. You can support my shows by signing up with Captivate here: https://www.captivate.fm/signup?ref=yzjiytz

==========================

About Jen

Jen's professional background is in web software technology (audio/video/web and graphics), working for many years in Silicon Valley. She has worked in instructional design, writing, marketing, and education in the creative space. She was also a quality engineer for awhile.

Jen became involved in performing, acting, and improv in 2015. She taught dance fitness classes (despite beginning with two left feet), performed in community theatre, and taught and coached improv comedy and acting at several theatres. Jen was also the Online School Director and Director of Marketing at WGIS.

Jen's website: https://jendehaan.com

This podcast is a StereoForest production. Made and produced in British Columbia, Canada.

Transcripts

WEBVTT

::

[MUSIC PLAYING]

::

So the internet is getting filled with a whole bunch

::

of AI-generated content.

::

And we're going to have to face this reality

::

because it's already well underway.

::

So we're going to need to adapt to this new internet

::

and figure out a way for the things that we

::

create to be found just to connect with other human beings

::

in the first place.

::

And if you're a creator or some kind of entrepreneur,

::

a good way to find connection is maybe with your flaws,

::

your personality--

::

not that that's a flaw--

::

but your personality and the way you think and your logic,

::

if it's impressive or kind of weird, your funny laugh,

::

the way you say things, whatever the things are

::

that you find joy in, all of these things,

::

whatever they are, they make you human.

::

Now, if you are one of these entrepreneurs or educators

::

creating content online, you've probably

::

seen how marketing is changing so much, so quick.

::

And one of the new things that a lot of marketers

::

are using are AI avatars that look surprisingly,

::

scarily human.

::

And these avatars are used because of the money that they

::

save.

::

You can put them online around the clock to sell your thing.

::

They look and they sound like really polished and perfect

::

or trendy and they never have bad hair.

::

And the metrics behind some of these virtual influencers

::

are pretty impressive, it seems.

::

When you look at it, I don't know if that's real, too.

::

Like, what is real anymore?

::

It's a pretty compelling idea if you're a marketing

::

and you're using something that can be efficient and

effective,

::

can just prompt whatever you want it to do.

::

And you can automate things to save time in addition

::

to this money saving.

::

But for those of us who believe in the power

::

of personal storytelling and genuine experiences,

::

you might be asking yourself, I'm asking myself,

::

like, what are some of the hidden costs here

::

of this perfection, this glossy look?

::

What's going to happen?

::

That, too.

::

So in this episode, I'm going to get

::

into the psychology of trust in the age of AI.

::

I will show you why our brains get kind of weirded out

::

by the things that are almost human and this concept

::

that you very well might have heard, Uncanny Valley.

::

I'm going to compare and contrast the Uncanny Valley

::

with another idea from psychology called the pratfall

effect.

::

And that explains why a person who's competent

::

can actually make a mistake and be more likeable somehow.

::

If you're lucky, it'll work for you.

::

But then I'll talk about what happens

::

when we mix those things together on Canny Valley

::

and the pratfall effect.

::

You're listening to the human internet theory,

::

and I'm your human host, Jen deHaan.

::

I started out in education and I've

::

worked in tech over, over a couple of decades,

::

most of it in web software in various capacities.

::

And I've also spent a decade in a side quest

::

as a performer on stages and digital in comedy.

::

And I'm talking about how to present things confidently

::

or in some way as a human in online spaces

::

and talk about some of these things that are coming up

::

and coming at us on the internet right now.

::

So before we talk about how we as human creators can do OK

::

or better or survive in this new landscape,

::

we need to understand the somewhat conflicting ways

::

our brains process perfection and imperfection.

::

So let's start with that feeling of unease,

::

that sense that something is just a little bit off.

::

Now, this isn't something new.

::

In 1970, a robotics professor at the Tokyo Institute

::

of Technology was called Masahiro Mori.

::

And Masahiro Mori wrote an essay about uncanny valley.

::

He hypothesized that as a robot looks and acts more human,

::

so as they progress towards humaneness,

::

our emotional response goes from neutral to positive.

::

And we feel empathy for that robot.

::

But however, as it gets almost perfectly human,

::

our response then completely changes

::

and it plunges from empathy down into a very negative space,

::

like revulsion.

::

And he called that plunge the uncanny valley.

::

And I'll link to the original work in the description.

::

Now, for a long time, that was just a theory, right?

::

But it's being studied scientifically.

::

So in a 2018 study, which I'll also link to,

::

it suggests that the uncanny valley reaction happens

::

because this robot or avatar violates our brain's

predictions.

::

So we see something that looks human.

::

So we expect it to move and behave in a very human way.

::

And when it doesn't, when the gestures are a little bit too

::

smooth or its eyes just don't quite track right,

::

it creates conflict in our brain.

::

And that conflict just feels kind of creepy or messed up

::

when we're looking at or hearing them

::

because it could be audio too.

::

For example, those 3,000 episodes every week

::

that an AI company plans to generate and put into our ears.

::

If what is being said or how it's being said

::

is uncanny valley that really close but just off vocally

::

through the cadence and all that kind of thing,

::

it could trigger this kind of response, a conflicted brain.

::

So our brains are wired to be suspicious of near perfection.

::

But contrasting the suspicion of near perfection

::

is that we often trust people more

::

because of their imperfections.

::

And this is explained by a concept in social psychology

::

called the pratfall effect.

::

So this was first written about by Elliot Aronson back in

1966.

::

So Aronson and his team studied how a simple mistake affects

::

how likable someone is.

::

And they found that when a very competent person makes

::

a small honest mistake like spilling some food

::

or making a mess or spelling something wrong,

::

maybe their attractiveness to other people actually goes

up.

::

But when a person who's perceived as average

::

makes the same mistake, their likability went down.

::

So according to this effect, a so-called like superior

::

or higher esteemed person becomes even more attractive

::

after that small and relatable blunder.

::

And now there's some added nuance around social

comparisons

::

and cognitive processing in a later study as well,

::

which I'll reference in the show notes so you can go check it

out.

::

But an idea to take away for the things that we're making

::

is that for someone generally to be respected,

::

a blemish of some kind can make them seem more human,

::

more approachable, more relatable, which is good for us.

::

So we have the uncanny valley where near perfection feels

wrong.

::

We have the pratfall effect where imperfections

::

can potentially build connections.

::

So what if we put these two concepts

::

alongside each other and see what happens

::

when we think about that?

::

Now, I saw something like this online recently.

::

So I follow creator in the social psychology space as well.

::

So I've followed this person for a while.

::

And I was watching one of their new videos

::

and it was like it was a technically impressive rendition

::

of their likeness.

::

Like I still had that feeling of something being kind of off.

::

Like I was like, oh, is this is this generated?

::

And but then near the end of the video,

::

a really big bug in the rendering caused this visual glitch

::

that made it really clear that what I was watching

::

was an AI generated avatar of the creator after all.

::

So my reaction to that was like just a bunch of questions,

::

and negative ones though.

::

If they weren't checking the final video for bugs like this,

::

are they checking the script?

::

Is the information that I'm hearing now

::

and we sort of set up this thing between us,

::

is that even accurate or real?

::

Does the creator even get involved

::

with making this piece of content?

::

Like any part of the process, topic, script, anything?

::

So in this case, the pratfall made me like that person's

::

content less honestly, which would relate to the social

::

competency thing if we go by the pratfall study.

::

So it seems it feels like it relates to that in addition

::

to the generated avatar that made me kind of uncanny valley.

::

At any rate, I was left with questions around integrity

::

and trust.

::

It's a little daunting to just think about where

::

these things are headed.

::

And if you work in or deliver educational topics,

::

it's particularly wild to think about what's

::

about to transpire here on the internet.

::

What's real and what's not in almost every category, like

::

scripts and avatars and everything.

::

So because we trust people or kind of know what's real or not

::

or what's put on because of these imperfections,

::

and because we often connect that way

::

through these imperfections, these honest mistakes.

::

Like from the beginning that I told about,

::

like the funny laugh or admitting to an error

::

or sort of stumbling over your own words

::

or the way you gesture like this or speak or when you're

sloppy.

::

These are all things that our brains have evolved over

thousands

::

of years to recognize as really human being relatable,

::

relating to each other.

::

And that's why the pratfall effect

::

is really a moment of relatability

::

that makes us more likable.

::

It depends on context and nuance, but in general,

::

it sort of puts us on the same page

::

and to relate to each other.

::

Like it puts us on the same page for that

::

so we can understand each other and humans like to be

understood.

::

All right, so what happens when the imperfection is made

::

by these AI avatars, like in my example that I gave?

::

Well, it's something different than the pratfall effect,

::

of course.

::

So before we get to the ways to differentiate ourselves

online,

::

we have to understand why we need to build trust online

::

for a human internet.

::

All of this really comes down to believability trust.

::

So when you make content and put it online,

::

there's kind of a parasocial relationship,

::

whether or not you want that.

::

There's positive and very negative parts

::

of parasocial relationships potentially.

::

But fundamentally, there is an exchange of information

::

being shared from person A to person B.

::

Now, an element of that can be broken

::

when something like the one thing I experienced happens,

::

which is a breaking of trust, kind of manipulation,

::

that kind of thing.

::

Because trust isn't really placed in the words being said

::

or written, and it's not being placed in the digital pixels

::

representing the human picture, audio, or text.

::

Trust is placed in the person, the human saying them,

::

or the human behind them.

::

And if that person is revealed to be a fabrication,

::

and it wasn't disclosed, that's an important point,

::

the audience might be wondering what is real

::

and what is not from that human source.

::

And our human brains look for that predictability

::

to know what is real or not.

::

And trust is built through many forms of cues,

::

through micro expressions, and voice, and cadence,

::

and rhythm, and these hand gestures.

::

So when your audience chooses to watch your content,

::

they're like casting a vote to invest

::

their most valuable asset, or at least probably one

::

fairly high on the list, which is their time.

::

Because time equals money and everything, whatever.

::

And it's finite.

::

And if you're educating, they might

::

be seeking certain things like building or wanting trust.

::

And as creators, we have a responsibility

::

to take that seriously if we want to build connections

::

for a human internet.

::

And results of polls do suggest that audiences

::

are leaning towards distrusting AI-generated marketing.

::

And this is referred to by some as the AI trust gap.

::

And I'll link to an article from MarTech

::

that speaks about this trust gap and some statistics

::

around it in the show notes.

::

And it mentions, among other stats,

::

that only 25% of customers like AI in email marketing,

::

70% of customers say they can recognize AI copy.

::

82% of customers want to talk to human agents,

::

even if the AI is more efficient.

::

60% of customers doubt the AI output, what it's putting out.

::

And 84% of audiences want disclosure

::

about this generated content.

::

So trust.

::

It takes a long time to build, and it can be lost in an instant.

::

And that one glitch.

::

And if you were listening to music in the late '80s--

::

so this is a real life example--

::

you might remember Milly Vanilly.

::

They built a huge career on some hit songs.

::

They won a Grammy.

::

And then they lost it all overnight

::

when a track of theirs started skipping

::

during a live performance.

::

So this was revealing that they were lip-syncing.

::

And, well, this pratfall, if you want to call it one,

::

was a deception, and their careers never recovered.

::

Now, the problem is that trust is fragile, of course.

::

That's the problem here that we're seeing.

::

But it's especially fragile in online spaces,

::

because we have such limited cues to begin with.

::

We don't have that in-person thing

::

that we can read a lot more off of.

::

And I'll link to a paper about deception specifically

::

in computer-mediated communication.

::

And this paper is a little bit old,

::

but it provides some really great insight in this space.

::

With the expansion growth of generated content,

::

it's really making reality hard to know what's real or not.

::

So I'd say this paper is especially relevant now.

::

And right now, the internet just feels really weird to me.

::

Like, when I see some things, I'm asking questions.

::

Like, was this a personal story or just a prompt

::

fed to a language model?

::

Did anyone live that experience?

::

Was the advice ever done?

::

Like, was it tested in the field?

::

Does the creator actually believe in this product

::

that they're recommending?

::

Are they recommending it?

::

Do-- is their endorsement also generated?

::

Like, do they even know that this video exists?

::

Or did they just check a box in some contract

::

to say their likeness could be used for whatever product

::

video, for whatever amount of money?

::

I mean, it's so fragile online.

::

And another question is, like, how many humans will care?

::

Or am I just some kind of dinosaur here,

::

not adjusting to the time?

::

Anyways, if you care about this stuff

::

and are finding this video useful,

::

head to my site at humaninternettheory.com.

::

You might like my newsletter too.

::

And you'll be able to get also a free three-page worksheet

::

that will help you build up your own real digital persona

::

that you're going to base off your own traits

::

so you can educate or inform your audience in a way

::

that you decide.

::

It's like about priorities and stuff.

::

Anyways, technology will advance.

::

I'm a dinosaur, but I'm going to come along with it.

::

And these avatars will get better,

::

maybe even indistinguishable from reality

::

and mostly glitch-free and are not too distant future.

::

But what happens then, right?

::

What is the human reaction when we're surrounded

::

by these like flawless, perfect, neat digital beings?

::

Now, my bet--

::

I'm going to bet on this--

::

we're going to crave realness and connection more than

ever,

::

even online.

::

Like, we're going to be looking for the offline third space

::

experiences, but we're still going to crave realness here.

::

And maybe especially online and especially

::

from sources where we're used to human-to-human

connection.

::

Now, maybe some of us might not mind those really big

::

capitalistic corporations using avatars and generated AI

::

because we don't have the same expectations from them

::

because we already don't trust those institutions.

::

However, that Mar-Tech article, check it out,

::

because they do suggest that maybe we won't

::

accept it from the large corporations either.

::

There's a case study in it.

::

But from the humans that we've connected with

individually,

::

the ones we rely on for those genuine sources of advice

::

or like real learned lived experiences,

::

I think in those spaces, we're going to crave those

pratfalls.

::

And we won't just want them from those superior people,

::

as they call them.

::

We're going to want them from anyone, I think,

::

that's left who is willing to be really genuinely human

online.

::

So in this potential future of the dead internet,

::

filled with bots and avatars and such,

::

the question for those of us choosing to build

::

on the human internet is this?

::

What can we focus on to differentiate ourselves?

::

And how can we maintain our social contract?

::

You know, thing that we have where we want to trust each other

::

and listen to each other.

::

How do we maintain that moving forward

::

amongst all of this generated content?

::

So first of all, one, we can embrace more unscripted

::

or semi-unscripted video.

::

Our unique imperfections, the things that make us human,

::

that is going to be our strength in this new internet.

::

So allowing for those spontaneous thoughts

::

off the top of the dome are natural pauses.

::

And finding your words in the moment

::

is a really good sign of authenticity.

::

And that might be a competitive advantage.

::

So this is really different than our history of marketing

::

that was so focused on these polished, perfect campaigns.

::

But online platforms, especially social media,

::

have really changed this over time.

::

There's more willingness to really engage between brands

::

and creators and educators and stuff

::

in a more genuine and transparent and sort of down-to-earth

::

not polished way.

::

It makes all of us more approachable.

::

And so if you do use scripts, you

::

can augment them with unscripted elements,

::

or you might move to notes instead.

::

You can add your additional thoughts in real time

::

along with what you're writing.

::

And this can be a learned thing to do,

::

sort of learning to speak off the cuff in a clean way.

::

You don't tangent or ramble too much.

::

It's something you can learn through improv classes.

::

But you can also practice this with reps just on your own.

::

While you make your videos iterate and improve over time.

::

One thing you can do is to remember to slow down

::

and gather your thoughts whenever you need to.

::

The pauses are fine.

::

And if you are editing your work afterwards,

::

as so many of us do, you can always cut really long pauses

::

out or cut it down to the right number of them.

::

You can also show some of the behind-the-scenes footage

::

or add behind-the-scenes stories specifically

::

into your content.

::

You can plan it out.

::

You can have that stuff with no scripts

::

and other stuff with scripts.

::

If you are a visual artist, you might

::

be doing some of those behind-the-scenes process,

::

show your process, the things that you make.

::

Now, needing to do this in the first place

::

is kind of a shame.

::

We didn't really need to do it before.

::

But it is the reality that we're heading into.

::

So create processes that help you make all of this easier.

::

You might also want to sort of get rid

::

of some of your lower priority work.

::

So you can prioritize the things that

::

do make more of a difference in your space.

::

If you're really into learning how

::

to do any of these things that I've mentioned in detail,

::

let me know in the comments if you're watching this on

YouTube

::

or send me an email if you're listening to the podcast.

::

Let me know what your main challenges are,

::

what you'd like to see more of.

::

And I will prioritize that content.

::

So second, understand that honest mistakes are

::

memorable too.

::

So don't over edit yourself.

::

Some of those things that you leave in

::

might be what make you stand out and stick in someone's head.

::

Leave a moment where you might break character and laugh

::

or have to correct yourself.

::

All of these little things can be a strength and a way to

connect.

::

And making mistakes and not beating yourself up for them

::

afterwards can be difficult if you're

::

prone to say listening to a very strong self-critic in your

head.

::

So a lot of this is about practice, those reps,

::

and confidence.

::

But more than anything, it's about mindset.

::

And I'll speak more about living with a strong self-critic

::

when you're making things in future episodes.

::

But I'll link to an episode I've already done on this topic

::

through the lens of improv performance in the show notes

::

if you want to check out something related today.

::

And finally, and this is a really important one,

::

disclose the use of generated AI visuals or audio.

::

And so if you're going to be using avatars of yourself

::

or AI-generated voiceovers, try to be transparent about

it.

::

It might help you in the long run.

::

Like turn on the toggle and the upload settings

::

on whatever platform.

::

You're often supposed to do it anyways to avoid penalty.

::

Put it in the description.

::

Make a little bumper for the front of your podcast

::

that says what it is before people start.

::

Put it in a burnt-in caption if you're doing a video.

::

So these forms of transparency when you combine them

::

with a really good review process to ensure quality

::

and accuracy is a good way to avoid--

::

might be a loss of trust if anything happens.

::

And people find out and it wasn't disclosed or whatever.

::

We don't know how the future is going to go

::

and what people are necessarily going to think.

::

But if people know what they're watching and it's very open,

::

this is what it is, there'll probably be a heck of a lot

::

more forgiving with whatever happens

::

if there's a glitch or something like the one that I saw.

::

Anyways, things are really changing

::

with all of these new features being added into tools

::

at very cheap or free or economical prices.

::

The allure of these perfect AI things,

::

it offers a very tempting path

::

because it's inexpensive and efficient and very polished

::

and kind of flexible to what you exactly want to say.

::

But it's also a path that does risk some things,

::

it sacrifices some really valuable things

::

that we have with our audience.

::

And that thing, like I've been talking about, is trust.

::

So we're gonna be flooded with these perfect avatars,

::

but our human imperfections, the stumbles

::

and how authentic we are

::

and these real unscripted moments that we might have,

::

those are a pretty good thing to keep.

::

So don't replace them.

::

Thanks so much for tuning in.

::

I'll be back next week with another episode.

::

(upbeat music)

::

(upbeat music)

::

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.

::

(upbeat music)

::

(upbeat music)

Links

Chapters

Video

More from YouTube