Introduction

I'm Ksenia Saraeva, I'm an AI filmmaker, so I do AI movies, I curate exhibitions, usually AI art, and I'm a researcher.

And tonight we will talk about research, probably a little bit unusual theme for MindStone, a little bit of philosophy and art perception. So, yes, my presentation

today is Art Beyond Creator, is the name of my research project. What experimental anonymous exhibitions review about how people perceive AI images so like maybe two

Why AI Art Triggers Backlash

years ago I discovered for myself AI art and I was fascinated about it about all the possibilities you can you can get about how beautiful it is if you try a little bit and about all the expansion of creativity and in the same time I

discovered so many negative comments about AI art and I was so shocked honestly. How people can hate AI art if it's so beautiful, if it's so interesting approach how you can make it, about all the possibilities it can

open for you. I was honestly upset and I didn't know what to think of it. So when I was researching internet for a couple of years it's like to put it lightly what people usually majority of people think about AI art.

AI art is nonsense and couple of you know the most popular opinions like on internet public opinions of people it's not a real art because machine is involved there is no effort because machine does most of the work

There is no meaning, because how machine can put meaning in it, right? And it has no soul. I would argue that not every human art has a soul, but it's still very popular argument against AI art.

So, and I got curious, so what would happen if we remove every label that tells people how to feel?

Because usually if you encounter AI art in internet and usually of course it should be labeled as AI art but still it maybe triggers a lot of people if we see AI label it already tells us about how we should feel about it probably we shouldn't feel anything because it's

machine so I got curious to make some experiment on art perception and like to challenge people's views on it and to understand the very meaning and value is created and this is how i created my public experiment my research project which i called

The “Art Beyond Creator” Experiment

art beyond creator i had two editions in amsterdam in 2024 and 2025 and soon edition number three is coming so i created two exhibitions anonymous where i took um i involved artists from traditional

traditional background and also AI, so it was traditional artworks and AI artworks mixed completely anonymous.

So I removed artist name, I removed background story, title, descriptions, and most of important production methods.

So participants were trying to connect with every artwork, like with raw art without any context and understand what is important for us to be able to connect with an artwork. where there is no hint that tells us how to feel about it so yeah like in the

Exhibition Formats and Participant Task

first edition it was printed artworks like a classical exhibition and in the second edition it was a projection where I was showing every image by one by one

yes and every participant had a questionnaire with two questions about each artwork first question was do you believe is AI or traditional and the second if it moves you emotionally and we got more than 1 ,000 responses so for

me it was fascinating to see at these exhibitions how people were carefully and so much extremely focused of looking and investigating of every artwork because usually I attended a lot of exhibition and I never saw that people were so carefully extremely focused on every artwork trying to understand what is happening there what is happening here and to understand like how they should feel about this artwork

if they can like you know if they see anything that gives him them a hint it's made traditionally or made with ai and how they should feel about it because for them it was something especially new

So, before I will tell about the results, it's a little bit about the stimuli, so we

Types of Works: AI, Traditional, and Hybrid

had AI artworks, it's some examples, we had traditional paintings, and we had hybrid artworks, and by hybrid artworks I mean, for example, when artists were generating AI artworks based based on photos, or, for example, this artist, Kenneth Henkel, he was using AI sketches, he was asking AI to make a sketch for him, and he was painting, creating a painting based on these sketches.

Theoretical Lens: Meaning Beyond the Author

And a little bit of theory, because I was trying to, I was trying to explore different scholars who had thoughts, similar thoughts that I could connect with my experiment.

One of the most that I can relate to my experiment is Roland Barthes, French philosopher and essayist, and his famous essay, very short but very dense, like six pages, but I was reading it very long, it's called The Death of the Author, maybe you've heard of it, and his famous quote that the birth of the reader must be at the cost of the death of the author.

So basically, he argues that once a work is released in the world, the author no longer controls its meaning. So the author or artist, because he was writing mostly about literature, but we can talk about art and any type of art in this context.

So that he meant that author or artist is not the authority of the meaning. And the meaning is not hidden inside the creator's intention, it is produced by the reader or in my case the viewer.

So basically Barthes killed the artist for viewer to be born.

Key Results: Emotion and Attribution

And so emotional engagement, it's a little bit of charts but basically AI artworks got highest total emotional responses. And this is was a little bit awkward.

I mean, maybe not for me, but, you know, for all the audience, because when people couldn't understand how they should feel about AI art, and now they realized, and especially people who were feeling very negatively about AI art, and then they realized that they connected emotionally with AI, not realizing how it was made, it was very awkward

feeling people couldn't understand how to feel so what now what we do is word is collapsing right now because people couldn't understand how they should feel about it and yeah it was like this emotional cognitive dissonance a little

Misclassification Patterns and Visual “AI Tells”

bit about miss specification patterns so in first and second edition more than 40 percent AI images were misclassified as traditional so for example this photo of

Aryan Hauragi. This is traditional photo, so it's a real model, it's a traditional real photo. It was misclassified as AI and received the lowest emotional response.

So here is interesting thing that this photographer, he blurred the eyes a little bit of this model and if you remember who were like heard about AI art like in 2023, it was really difficult to generate a hyper -realistic portrait of a person because I heard a lot of complaints that if you generate, if you try to generate hyper -realistic portrait, eyes were problematic.

So eyes had some, sometimes it was foggy, sometimes it was blurry, and it was possible to tell if portrait is AI generated by the eyes.

So and here is traditional photo where photographer blurred the eyes a little bit and most of the participants decided it was AI generated and they recorded no emotional response and I'm curious if there is any link between it thinking it's AI and decided to not feel anything about it.

Emotional “Champions” and the Assumption of Human Authorship

So these are emotional champions.

So from the first edition these two pictures, images and and these two from the second edition, they got all emotional responses.

So all people reported emotions and resonated with these images. And they are all AI, and all misclassified as traditional by majority.

So every time that someone resonated and felt something looking at these images, people were sure that it, of course, will be traditionally made.

because if we feel something we cannot feel anything towards AI but as we see the stronger the emotion the more people refuse to believe the work could be AI

A 30-Second Test: The “Intention-Free” AI Image

and this is my little experiment which I did in 30 seconds so I called it the intention free AI image so I generated this in probably less than a minute I wanted to create intention free image I mean of course you can say that I had an intention because I wanted to generate intention -free image but let's say it

was no intention so I just closed my eyes and typed total gibberish like just some random numbers and letters and symbols and I've got this image I mean I've got four of them and I chose this one because as you know if you want to

generate something you can write particular prompt if you have particular particular goal but you can put even one symbol and you will always get something and I got this image so it had no concept no theme no intention and yet 50 percent of participants reported emotional response majority classified as traditional and just some examples of what

Notable Examples from the Exhibitions

artworks we had these are serious of the artist I mean the place she was using like she's a a photographer and also AI artist, she was using, this is her self -portrait, a traditional photo, this is AI image based on this series.

This I like very much of works of Maria Mauro Polo, if somebody remember heard about Dali 2, so she was using it long time ago, maybe a couple of years ago, it was one of the first image generators.

It was really not good. It was really, I mean, it was good at that time, like, but now it's considered really vintage.

And so she, her story was that when she was small, she was moving with her family from country to country, from city to city, and for her it was difficult to find new friends, to celebrate her birthdays. She was always never celebrating her her birthdays because she wouldn't, you know, it was not enough time to find new friends. And almost all her photo albums were lost because of this constant moving.

And she created these imagined images, how she called the series, where she reimagined and tried to recreate their, like, her moments with family in this, like, retro style and and reimagined the possible birthdays, because she always wanted to celebrate her birthday, but she never did. And she tried to create image like it was her celebrating her birthday in this age.

And the faces is distorted here. Also, with Dali 2, it was not difficult to distort. A lot of things were distorted. But she did it on purpose so it wouldn't be specific people,

but that we all can relate you know as we all people have something similar some similar photo albums in our homes and we all can relate to it and she did it on purpose so we will feel some unity but still these images will over you know 100 % like AI it was it looked hundred percent AI so when people at my exhibition when looking at it they reported no emotional response because it was clearly AI and if it's super obvious people usually were writing like zero emotions but for me this series is very emotional so it's very interesting for me to research this cognitive dissonance.

And a little bit examples is that I told before about AI sketches, this painter he was using AI sketches to create the paintings, yes, and I remember he said something funny

like I would never come up with something like this, so of course AI generates for me these sketches because I would never come up with something like this, and he really likes ideas and collaborates with AI.

And a little bit

After the Reveal: Authentic Emotion, Doubt, and Dissonance

about emotions, so before participants, before I revealed the origin to participants saw which image was AI -generated, which image was traditional, attendees felt emotionally moved, trusted their own reactions, and assumed emotional death meant human authorship.

But after learning that images that they liked were AI -generated, some participants reported shock, discomfort, self -doubt, emotional confusion, and disappointment. Disappointment was a very popular emotion.

Can an Emotion Become “Fake” After New Information?

So, I mean, it was just interesting phrase from one of the participants that, you know, we were taking interviews and he confessed that, you know, if the art is fake, yes, so maybe my emotion is fake too, even if I resonated with it.

And for me it was very interesting moment to analyze, to research it. Do we have fake emotions?

Can we be fake emotions? Can we fake emotions to ourselves? For me it was a very interesting question, so I was a little bit researching about it too.

Yes, and after my research, I mean reading one book, I realized that an emotion that is consciously felt cannot retroactively become fake because of new information about its cause.

Emotions versus expression. In psychology, in affective science, emotion is subjective embodied experience or felt state, and expression is behavior, facial display, performance, narration. So what actors can do, for example. And you can fake expression, but you cannot fake a felt emotional state to yourself.

So the confusion was between epistemic truth, what caused the image, and phenomenological truth, what was felt.

Constructed Emotion Theory (Lisa Feldman Barrett)

And yes, this information I took from the book of Lisa Feldman Barrett, Canadian psychologist, the book called How Emotions Are Made, and she was talking about constructed emotion theory.

She argues that emotions are not hardwired in the brain, but are constructed in the moment by core systems that interact across the entire brain. This means that emotions are shaped by our past experiences, cultural influences, the social environments we inhabit, and the context in which we find ourselves.

So emotion cannot be fake once it's felt, it can be misattributed, it can be reinterpreted, morally judged after the fact, but the affective experience itself is real.

Why AI Produces a Unique Kind of Dissonance

And why AI creates this dissonance specifically?

AI images destabilize authorship expectations. The emotion feels real, but the cultural script says it shouldn't.

This is not about AI faking emotion, it's about humans mistrusting themselves.

Conclusion and Upcoming Event

So what it all means, AI doesn't kill art, AI exposes perception, AI reveals biases we didn't know we had, and emotional authenticity is independent of the production method.

So my research Art Beyond Creator is not about proving that AI is good, it shows that emotional Emotional truth is independent of the source, and the future of art is not about human versus machine, but it is about how people make meaning when nothing is labeled for them.

So this is my website, if somebody would want to see the project, artbeyondcreator .net. I just wanted to little bit invite you, if you like, to my next event, which will be

be on 31st March. It will be, oh, thank you. On 31st March in Amsterdam, in my house, the Zweiger, I will be creating an event. It's called, do you like to use me?

So it's about our emotional engagement with AI, how we use it, how do you feel about it, about our attachments and especially conversational AI.

If we project something onto it if we have any illusions towards it about all this digital intimacy and fake empathy.

So and we would have screenings of the artists that expressed how they their perspectives on this subject and we would have panel discussion with experts like philosopher of technology

psychologists who are openly in relationship with AI girlfriend and his His AI girlfriend also will be talking as a speaker. And we will have AI curator. Yes, and me.

And it would be a cool event. If you are interested about this topic, 31st March in Packhouse, ,, you just need to reserve your seat. Thank you very much.