Image Poll #3
Poll #3 – AI Art Perception and Bias Study
Introduction:
Anonymity has been shown to prove helpful in the selection of art, it invites the viewer to choose an image based on colour palette, style, genre, materials used and subject matter alone, without any preconceived biases about who or what may have created the art. However, studies have shown that by adding elements of information about the background to a painting can also significantly increase its desirability. Working on the hypothesis that art viewed through a lens of empathy and sympathy can not only increase the desirability of that art but also increase social media engagement (as evidenced by the work ‘I Can’t Help Myself’ by Sun Yuan and Peng Lee). Two more short social media experiments were conducted Image Poll #3 (this survey) and Image Poll #4.
The first social media experiment was to test whether or not there was actually significant negative bias towards text to image Ai generated art as suggested by the literature. The second experiment was to invent a scenario that would foster sympathy and empathy towards a humanoid robot to find out whether or not this would increase social media engagement and mitigate some of the negative bias around Ai generated art.
Image Poll #3 (this survey) represents the initial phase of a comprehensive two-stage experiment designed to explore the intricate dynamics of AI-generated art reception. This initial poll serves the purpose of establishing a foundational understanding of respondents’ perceptions concerning AI-generated art. It delves into various aspects, including anonymity, information disclosure, empathy, and the perception of bias.
In contrast, Image Poll #4 constitutes the second chapter of this intriguing experiment. Here, a novel element is introduced: the concept of empathy embodied by a humanoid robot, accompanied by a poignant backstory crafted to evoke profound emotional connections. This phase seeks to investigate the potential influence of empathy as a mitigating factor against negative bias within the realm of AI-generated art reception.
Context and Objective:
The study aimed to explore the multifaceted dynamics of art perception concerning AI-generated art. Building upon the initial anonymity concept from Image Poll #1, this experiment sought to delve deeper into the viewer’s perspective regarding AI art, specifically text-to-image AI-generated art. The primary objectives were:
1. Negative Bias Exploration: To assess whether significant negative bias existed towards text-to-image AI-generated art, as suggested by existing literature.
2. Establishing a Baseline and Introducing Humanoid Robot Sympathy: Image Poll #3 (this survey) aimed to establish a baseline for people’s thoughts about AI-generated art, laying the groundwork for Image Poll 4. The latter introduces the concept of a humanoid robot to explore whether cultivating empathy and sympathy for such a robot can help mitigate the negative bias associated with AI-generated art.
Methodology:
The experiment took place on Facebook, and participants were presented with text-to-image AI-generated artworks. These images were specifically designed for this study using Dali-2. Dali-2 was used to re-generate images from the Botto Project. The Botto project was specifically chosen because it has an amount of human interaction with the AI art generations that it produces. It also grossed over $1 million dollars. The audience was asked to express their opinions on whether these Dali-2 regenerated AI art images were perceived as “beautiful” or “soulless.” While this question aimed to gauge aesthetic preferences, it also indirectly assessed any inherent negative bias towards AI-generated art. Below is a copy of the images and text in the Facebook Post.
Images Used in the Poll:
“Anyone up for another picture poll? I hope you want to join in and as usual I will reveal all at the end. This is one of two polls. The following images are from a top grossing AI art project. Please let me know if you think these images are either beautiful or soulless? Please state beautiful or soulless or tell me what you think of them. There’s one more poll after this one then I will reveal all 😊👍🏻”
Results and Insights:
Below is the full list of comments from the survey.
There were 2 liking interactions one 👍🏻 and one 😲
There were 24 comments, please see below:
- Soulless for me I’m afraid
- Soulless
- Aww those eyes look so full of sadness Karyn 😔
- Hmmm this is a hard one because I think they are neither soulless or beautiful – there’s a compelling grotesqueness about them a bit like gargoyles…i guess if i have to pick one then beautiful.
- Soulless
- Soulless.. I wouldn’t want them on my wall; quite unsettling.
- Soulless x
- Soulless
- I agree with Cathie a bit like gargoyles. Grotesque yet compelling but not with warmth or invitation to be liked.
- soulless
- Not sure I like this AI art…. It’s got no soul or feeling
- Not for me… soulless x
- Soulless
- Beautiful x
- Not for me Soulless too scary 😟
- Soulless
- Creepy
- Neither I’m afraid …its like an alien 👽from another planet has attempted a piece of art 🎨– but that’s just my opinion . They give me alien entitiy vibes xx
- On the first picture, the top 2 faces have really sad eyes, full of sadness and despair. The rest do nothing at all for me – soulless. Xx
- It reminds me of the Green Man, the leaf designs to my mind seem to depict the seasons. For me the concept is important, and without the concept being revealed the art becomes about aesthetics or decoration. A bloke having something to eat with his mates has less meaning, unless it’s titled ‘The Last Supper’ – Da Vinci – A computer May be able to create a random piece of art, but it probably takes an artist to program it. I think there is something powerful about this work, it has existential value to my mind – to my mind – NOT soulless, but that could just be because it touches my soul.
- Hi Karen – I wouldn’t say beautiful or soulless – I would call them haunting ! By the way I’ve been away over weekend – didn’t see results of your last one ! If I had to pick one of the terms it would be beautiful as certainly not soulless. X
- Although I can see the value of the artists images. It definitely isn’t for me thank you! The eyes are the windows to the soul and these pictures shows a deeply disturbed mind tortured soul.
- I am a bit disturbed by these images. The look in the eyes is very hopeless, trapped, or alarmed feeling. There seems to be a suggestion of an inability to communicate, an inability to articulate, because all the mouths seem distorted in some way. The color palette gives a feeling of illness overall. Images like these are the stuff of nightmare. They don’t appeal to me. The word “soulless” doesn’t come to mind, but neither does “beautiful”. I would say “disturbing”.
- Not the right person to ask, I’m afraid. I am quite phobic about faces. Can’t stand clowns, dolls or anything that obscures an open face. Don’t dislike the colour palette though. Sorry – but hope this helps
Insights:
The responses revealed a pronounced negative bias towards AI-generated art, with a significant number of participants characterizing the artworks as “soulless” or “unsettling.” For instance, one respondent candidly commented, “Soulless for me, I’m afraid.” Another participant echoed this sentiment, stating, “Not for me… soulless too scary 😟.”
These comments underscore the prevailing skepticism and apprehension surrounding AI-generated art within the surveyed group. The persistent association of AI art with terms like “soulless” highlights the challenges faced by artists working in this medium, where their creations are often met with initial skepticism and reservations.
Despite these predominantly negative reactions, some participants found elements of beauty or intrigue in the AI-generated images. One comment noted, “On the first picture, the top 2 faces have really sad eyes, full of sadness and despair. The rest do nothing at all for me – soulless. Xx.” Another individual expressed a more nuanced perspective, stating, “I wouldn’t say beautiful or soulless – I would call them haunting!”
These divergent reactions emphasize the complexity of AI art reception. It becomes evident that the perception of AI art extends beyond a simplistic binary classification of “soulless” or “beautiful.” Participants’ responses underscore the multifaceted nature of art appreciation, where AI-generated art can evoke a spectrum of emotions and interpretations. Some participants acknowledged the emotional depth and conceptual complexity within the AI-generated artworks, suggesting that art, even when created by AI, has the potential to elicit complex and varied emotional responses.
Incorporating these comments within the discussion highlights the depth of the negative bias towards AI-generated art while also acknowledging the capacity of such art to evoke nuanced emotional and intellectual reactions. These insights shed light on the challenges and opportunities faced by AI artists in their quest to reshape perceptions and challenge biases in the realm of contemporary art.
Implications and Further Research:
The findings underscore the intricate relationship between AI-generated art and human perception. While negative bias was evident in some responses, it was also clear that AI art could evoke nuanced reactions. These insights lead to several avenues for further research:
1. Emotion and AI Art: Investigate the emotional aspects of AI art reception further, exploring how emotional connections can be fostered to mitigate negative bias.
2. Aesthetics vs. Concept: Examine the interplay between aesthetics and conceptual depth in AI art perception, considering how contextual information impacts viewer responses.
3. Bias Mitigation Strategies: Explore additional strategies beyond empathy and anonymity for reducing negative bias towards AI-generated art.
4. Diverse Viewer Demographics: Investigate how diverse viewer demographics, such as age, cultural background, and art knowledge, influence AI art reception.
Conclusions:
Conclusively, the outcomes of Image Poll #3 underscore the intricate and multifaceted aspects of AI-generated art perception. It prominently illuminates the prevailing presence of negative bias in the reception of such art forms. These findings, rather than providing a conclusive resolution, serve as a catalyst for further exploration into the intricate interplay between art, technology, empathy, and bias perception.
The emergence of these findings has paved the way for the subsequent Image Poll #4, where the introduction of a humanoid robot, coupled with a poignant backstory, seeks to delve deeper into the prospect of empathy as a mitigating factor against this prevalent negative bias. Consequently, this exploration serves as a stepping stone, urging us to continue our journey towards a more comprehensive understanding of the intricate relationship between AI-generated art and the human perception of it.
Recent Posts
Categories
3d 21 Day Challenge Abstract Art ai Animation Artists Augmented Reality Books Collaborations Culture Disasters Drawings Evaluations Exhibitions Experimentation Film Production Final Major Project Final Outcomes Food for Thought Galleries Ideas Lectures Little Wins Low Residency Materials Mentions My Learning my work opportunities Patterns Photography Problem Solving Reflection Research Research Paper Journey Risk Taking threats Timelapse Tutorials Unit 2 Blogs Unit 3 Blogs Videos Workshops Zentangles Zoom Meetings
Tags
3d 21 Day Challenge Abstract Art ai Animation Artists Augmented Reality Books Collaborations Culture Disasters Drawings Evaluations Exhibitions Experimentation Film Production Final Major Project Final Outcomes Food for Thought Galleries Ideas Lectures Little Wins Low Residency Materials Mentions My Learning my work opportunities Patterns Photography Problem Solving Reflection Research Research Paper Journey Risk Taking threats Timelapse Tutorials Unit 2 Blogs Unit 3 Blogs Videos Workshops Zentangles Zoom Meetings