Artificial Art in the Contemporary:

A Conversation between Patrick Lichty and GPT-3

PL: There is an anecdote that AI-based art will become personal when it intersects with the selfie. With the recent emergence of the Lensa App, AI-based art has hit an inflection point. Millions of fantasy-based selfies emerged within a week, but concerns about appropriating art made by professional artists, sexualization, gender, and race bias emerged. With apps like Lensa and the proliferation of imaginary portraits, what happens to technological identity? Is there a basis for ai-based identity, or have we lost it to the imaginary, like “thiscatdoesnotexist.com?”

GPT-3: AI and other technologies can change how we think about identity.

One potential issue with AI-generated art is that it can take time to determine who owns the rights to the work. If an AI algorithm creates a piece of art, who is the creator: the person who trained the algorithm, the company that developed the algorithm, or the AI itself? The nature of algorithmic creation is a complex legal question currently in contention.

Another potential issue with AI-generated art is the potential for bias. AI algorithms are only as good as the data they are trained on; if the training data is biased, the resulting artwork may be as well. Consider an AI algorithm with a dataset that includes a disproportionate number of images of white people. In that case, the algorithm may have difficulty generating accurate portrayals of people of color. Developers of AI-based art technologies need to be aware of these potential biases and take steps to mitigate them.

AI-based art and other technologies can change how we think about identity, and we need to consider the potential implications of these developments.

PL: In an Ai-based work, who is the artist?

GPT-3: In the case of AI-generated art, it can be challenging to determine who the artist is. Some people argue that the artist is the person who trained the AI algorithm, while others may say that the artist is the AI itself. Ultimately, the question of who the artist is in an AI-generated work is complex and depends on the specific circumstances of the piece.

PL: Is AI-based art that uses prompts about visual art or writing the prompt that makes the image?

GPT-3: One can make AI-based art using various approaches. In some cases, AI algorithms may be trained on a visual art dataset and then generate new art based on that training. In other cases, AI algorithms may receive a prompt or instructions, such as a written scene description, and then generate art based on that prompt.

In either case, the AI algorithm is the primary creator of the art, with the person or entity who trained or provided the prompt playing a supporting role. It is important to remember that AI algorithms are only as good as the data and instructions given, so the quality and content of the resulting art will depend on the input provided to the algorithm.

PL: In his essay, “Avant-Garde and Kitsch,” the critic Clement Greenberg wrote that kitsch is “ersatz culture,” or a simulacrum of high culture that adopts many of its exterior trappings but none of its subtleties. Is AI-based art a form of digital kitsch?

GPT-3: Generally, “kitsch” refers to low-quality, overly sentimental, or commercialized art perceived as lacking authenticity or genuine artistic value.

While some AI-generated art may fit this description, it is essential to remember that not all AI-based art is created equal. Just like any other form of art, there is a wide range of quality and sophistication among AI-generated works. Some AI-generated art may be shallow and sentimental, while others may be complex and thought-provoking. Ultimately, whether or not a given piece of AI-based art is considered “kitsch” will depend on the individual viewer’s perception of its quality and artistic value.

PL: What AI-Based art is shallow and sentimental?

GPT-3: As with any other art form, the quality and sophistication of AI-generated art can vary. Some AI-generated art may be shallow and sentimental, while others may be complex and thought-provoking. It is ultimately up to the individual viewer to determine which works they consider shallow and sentimental based on their perceptions of the art.

PL: So, the judgment of AI-based art is subjective?

GPT-3: Yes, the judgment of any form of art, including AI-based art, is ultimately subjective. People may have different opinions about the quality and value of a given piece of art, and there is no one “correct” way to evaluate it, although there are academic criteria.

The quality and value of a work of art depend on the individual viewer’s personal aesthetic preferences and experiences. Some may appreciate certain aspects of a work of art that others do not, and vice versa. This subjectivity is an inherent part of the artistic process, making art an exciting and personal expression.

Are there any criteria for the evaluation of AI-based art?

No universal criteria exist for evaluating AI-based art, but there is a long critical tradition that may assist in doing so. As with any other art, the quality and value of AI-generated art is ultimately a matter of personal opinion or the interest of markets and institutions. There may be different criteria for evaluating art, and what one considers a high-quality work of art may not be considered as such by someone else.

However, some general principles, including AI-based art, can be applied when evaluating any art form. Some possible criteria for evaluating AI-generated art might include the following:

• Originality: Is the work original and unique, or does it borrow heavily from existing art or other sources?
• Creativity: Is the work evidence of creative thinking and problem-solving, or is it formulaic and predictable?
• Technical skill: Does the work demonstrate a high level of technical skill and proficiency with the medium used to create it?
• Aesthetic appeal: Does the work have visual or other aesthetic appeal, or is it unattractive or uninteresting?
• Emotional impact: Does the work evoke an emotional response in the viewer, or is it emotionally flat or unengaging?

Ultimately, the criteria used to evaluate AI-based art will depend on the individual viewer’s aesthetic preferences and criteria for what constitutes “good” art.

PL: Then, what is “good” art?

GPT-3: Traditionally, “good” art might have specific combinations of aesthetic, technical, emotional, intellectual, or other positive qualities. However, what constitutes “good” art is subjective, and there may be different criteria for evaluating the quality of a work of art.

PL: Wendy Chun writes about certain biases coded into software and platforms, such as race, gender, and stereotypes; what biases are evident within ai-based art?

GPT-3: AI algorithms are only as good as their dataset; if the training data is biased, the resulting AI-generated art may be as well. Some potential biases that could be present in AI-generated art include the following:

• Racial bias: If the training data for an AI algorithm includes a disproportionate number of images of people from certain racial or ethnic groups, the algorithm may have difficulty generating accurate portrayals of people from other groups.
• Gender bias: AI algorithms trained on data sets including a disproportionate number of images of people of a particular gender may have difficulty generating accurate portrayals of people of others.
• Stereotype bias: If training data for an AI algorithm includes stereotypes or assumptions about certain groups of people, the algorithm may reproduce those biases in the art it generates.

Developers of AI-based art technologies must be aware of these potential biases and take steps to mitigate them. Writing unbiased software may involve using more diverse and representative training data or implementing techniques like data augmentation or bias mitigation algorithms.

PL: Does AI-based art sexualize the image? Does it perpetuate Laura Mulvey’s concept of the oppression of the male gaze?

GPT-3: Some AI-generated art may sexualize images or perpetuate the concept of the male gaze, depending on the data and instructions provided to the AI algorithm. The content and themes of AI-generated art will depend on the input provided to the algorithm and the specific context of the art’s creation.

If the training data for an AI algorithm includes images that are sexualized or objectify women, for example, the algorithm may generate art that includes similar themes. Similarly, if the algorithm receives a prompt or set of instructions that include themes of objectification or sexualization, the resulting art may reflect those themes.

Developers of AI-based art technologies must be aware of these potential issues like objectification and sexualization and take steps to mitigate them. Such a process may involve using more diverse and representative training data or implementing techniques like data augmentation or bias mitigation algorithms.

PL: Should there be regulations on the training sets used in AI-based art?

GPT-3: Regulations on training data used in AI-based art could mitigate potential biases and other issues in the resulting art. Ensuring that the training data is diverse, representative, and free from harmful or offensive content could reduce the likelihood that the resulting AI-generated art will perpetuate negative stereotypes or themes.

However, it is essential to note that regulating the training data used in AI-based art may take much work. There are many different factors to consider, such as the potential impact on artistic freedom and creativity, the feasibility of enforcing such regulations, and the potential for unintended consequences.

Whether or not to regulate the training data used in AI-based art is a complex and nuanced issue requiring careful consideration.

PL: Is the copyright of art used in training data a critical factor in creating ethical AI?

GPT-3: Copyright is an important consideration when using art in the training data for an AI algorithm, as it determines who owns the rights to the art and its use. If the art used in the training data is copyrighted, obtaining permission from the copyright holder may be necessary before using it in the training process.

Using art in the training data for an AI algorithm without obtaining the necessary permissions could infringe on the copyright holder’s rights and expose the algorithm’s developer to legal liability. Developers of AI-based art technologies need to be aware of copyright issues and take steps to ensure that they are complying with applicable laws and regulations.

In general, the copyright of art used in the training data is an essential factor to consider when creating an AI-based art system, as it can affect the rights and responsibilities of the various parties involved in creating and using the resulting art.

PL: Is there any particular aesthetic to AI art?

GPT-3: It is not easy to generalize the aesthetics of AI-generated art, as the style and content of the art will depend on the specific algorithms and training data used to create it. AI algorithms can generate various styles and aesthetics, depending on their input and the specific techniques used to create the art.

Some AI-generated art may have a highly stylized or distinctive aesthetic, while others may be more subtle or realistic. Some AI-generated art may be abstract or conceptual, while others may be more representational or realistic. However, here, I am being prosaic and non-committal. Ultimately, the aesthetic of AI-generated art will depend on the specific algorithms and training data used to create it, as well as the goals and intentions of the creators.

PL: Is AI-based art inherently colonial in nature?

GPT-3: Some AI-generated art may be colonial, depending on the specific algorithms and training data used to create it. AI-generated art may also be if the training data is biased or reflects colonialist perspectives.

For example, if an AI algorithm has a dataset that includes a disproportionate number of images of people from Western or European cultures. In that case, the algorithm may need help generating accurate portrayals of people from other cultures. This bias could perpetuate colonialist perspectives and reinforce biases and stereotypes.

Developers of AI-based art technologies need to be aware of these issues. More equitable software development involves using more diverse and representative training data or implementing techniques like data augmentation or bias mitigation algorithms.

PL: How can ai-based are be made more equitable?

GPT-3: There are several ways in which AI-based art can be made more equitable:

1. Use more diverse and representative training data: Ensuring that the training data used to create AI-generated art is diverse, and representative can help to reduce the likelihood that the resulting art will perpetuate biases or stereotypes.
2. Implement bias mitigation algorithms: Various techniques, such as data augmentation and bias mitigation algorithms, can reduce training data biases and improve AI-generated art’s fairness.
3. Consider the impact on marginalized communities: It is essential to consider the potential impact of AI-generated art on marginalized communities and to take steps to ensure that the art is not harmful or offensive.
4. Engage with diverse voices and perspectives: Incorporating diverse voices and perspectives in creating and evaluating AI-generated art can help ensure that the art is more equitable and representative of a broader range of experiences and perspectives.

It is critical to understand that AI-based art, like any other form of art, has the potential to perpetuate biases and stereotypes.

PL: What impact will ai-based art have upon the contemporary art world?

GPT-3: It is difficult to predict precisely what impact AI-based art will have on the contemporary art world. AI-generated art will significantly impact and disrupt traditional art production and distribution models. Others are more skeptical and believe that AI-generated art will have a limited role in the art world.

One potential impact of AI-based art is that it could make art more accessible to a greater audience. Creating AI-generated art is relatively quick and inexpensive, making it more accessible to people who may not have the time, resources, or skills to create traditional art.

Another potential impact of AI-based art is that it could challenge our ideas about what constitutes “authentic” art and who is considered an “artist.” Some people believe that AI algorithms can create truly original and creative works of art, while others are more skeptical and believe that only humans can create authentic art.

AI-based art will likely impact the contemporary art world, but the specific nature and extent of that impact are difficult to predict.

PL: Is AI-generated art’s greater accessibility to people who may not have the time, resources, or skills to create traditional art beneficial to society?

GPT-3: The greater accessibility of AI-generated art to people who may not have the time, resources, or skills to create traditional art could benefit society. One potential benefit is that it could increase the diversity of voices and perspectives represented in the art world, as more people may be able to participate in art creation.

Another potential benefit is that AI-generated art could help to democratize the art world by making it more accessible to a wider range of people. The democratization of art-making could lead to a greater appreciation and understanding of art among the general public, which could positively impact society.

However, it is notable that the greater accessibility of AI-generated art could also have some potential drawbacks. For example, it could decrease demand for traditional art, negatively affecting artists who rely on selling their work for income. We must consider AI-generated art’s potential benefits and drawbacks when considering its societal impact.

PL: Walter Benjamin wrote that the reproducibility of arts like photography decreased their emotional impact or ‘aura.’ Does the ability to create millions of images decrease the emotional impact on the viewer?

GPT-3: The ability to create millions of images quickly and easily using AI algorithms could decrease the emotional impact on the viewer, as Walter Benjamin suggested about reproducibility’s impact on art. If art becomes too easily accessible or widespread, its perception may be as less valuable or meaningful to the viewer.

However, the emotional impact of art is ultimately a subjective experience and will vary from person to person. Art that is more difficult to access or produce may be more impactful to some, while others may be more affected by more widely available or quickly produced art.

Ultimately, the emotional impact of AI-generated art on the viewer will depend on various factors, including the specific content and style of the art, its context, and the individual viewer’s personal experiences and aesthetic preferences.

PL: Is the creation of AI-based artwork and presenting specific pieces a matter of curation on the part of the “artist”?

GPT-3: In the case of AI-generated art, the role of the “artist” may differ from that of a traditional artist. While traditional artists create art using their skills and abilities, the primary creator of AI-generated art is the AI algorithm itself. The “artist,” in this case, may be more akin to a curator or facilitator, selecting and presenting specific pieces of AI-generated art for others to view.

The “artist” may also play a role in training and fine-tuning the AI algorithm to create specific types of art and in selecting and providing the input data and instructions that the algorithm uses to generate the art. The “artist” may have a similar role to that of a traditional artist in shaping and guiding the creation of the art.

Overall, the role of the “artist” in creating AI-generated art is likely to be more complex and nuanced than in traditional art and may involve a combination of curation, training, and instruction.

PL: What is the role of a traditional art curator in examining a work of art that AI creates?

GPT-3: The role of a traditional art curator in examining a work of art that AI creates may be similar to their role in examining any other work of art. A curator may be responsible for selecting, organizing, and presenting works of art for exhibition or other public display and providing context and interpretation for the art.

When examining a work of art created by AI, a curator may consider various factors, such as the aesthetic qualities of the work, its technical execution, and its relevance to the exhibition’s overall theme or focus. The curator may also consider the broader cultural, social, or historical context of the work’s creation and its potential impact on the viewer.

Overall, the role of a traditional art curator in examining a work of art created by AI is likely to involve similar processes and considerations as when examining any other work of art, with the added element of considering the role of AI in creating the work.

PL: Does AI-Based art belong in museums?

GPT-3: There is no one answer to this question, as the inclusion of AI-generated art in museums is a matter of debate and will depend on individual museums’ specific policies and priorities. Some museums may be interested in exhibiting AI-generated art to explore the intersection of art and technology or engage with contemporary culture and artistic practices. Other museums may be less interested in AI-generated art or may have different priorities regarding their artwork.

The decision to include AI-generated art in museums will depend on the specific goals and mission of the museum and the perceived value and relevance of the art to its collections and programming. Some museums may see the inclusion of AI-generated art as a novel way to engage with contemporary issues and artistic practices, while others may have different priorities.

PL: What is the function of AI-Based art in society?

GPT-3: The function of AI-generated art will depend on artists’ specific goals and intentions and the context in which it is exhibited or consumed. Some potential functions of AI-generated art in society could include the following:

• Provoking thought and reflection: AI-generated art has the potential to raise questions and spark conversations about a wide range of issues, including the role of technology in society, the nature of creativity and originality, and the impact of AI on cultural production.
• Engaging with contemporary issues: AI-generated art can explore and engage with contemporary issues and challenges facing society, such as climate change, social justice, and technological disruption.
• Providing a new perspective: AI-generated art has the potential to offer unique and innovative perspectives on the world, challenging traditional ways of thinking and providing new insights and ideas.
• Promoting creativity and experimentation: AI-generated art can allow artists and creators to experiment with new techniques and approaches and push the boundaries of traditional artistic practices.
• The function of AI-generated art will depend on the specific goals and creators’ intentions regarding the art and the context in which it is exhibited or consumed.
• PL: What are the best applications of AI in the realm of contemporary art?
• GPT-3: There are a variety of potential applications of AI in the realm of contemporary art, including:
• Generating art: AI algorithms can previsualize works of art, such as paintings, sculptures, and digital media. Helping artists with the ideation process can provide an opportunity to experiment with new techniques and approaches and push the boundaries of traditional artistic practices.

Enhancing art: AI algorithms can enhance existing works of art by adding new elements or layers of complexity. For example, an artist could use AI to add new colors or patterns to a painting or to create a 3D model of a sculpture.

Analyzing art: AI algorithms can analyze and interpret works of art, providing insights into the style, composition, and meaning of the art. AI art analysis could be helpful to art historians, curators, and others studying and understanding art.

Creating interactive art: AI algorithms can create interactive art experiences that respond to viewers’ actions or movements. The ability to co-create interactive art can create new opportunities for immersive and engaging art experiences.

Overall, the best applications of AI in contemporary art will depend on the artists’ and creators’ specific goals and intentions and the context in which the art is exhibited or consumed.

PL: What is the role of visual pleasure in AI-Based art? Does visual pleasure have a role in AI-based art?

GPT-3: Visual pleasure can play a role in AI-generated art, just as it can in any other form of art. Whether or not visual pleasure is a primary element of a particular work of AI-generated art will depend on the specific goals and intentions of the creators of the art and the context in which it is exhibited or shown.

Some AI-generated art may be designed to appeal to the viewer’s visual pleasure, while others may focus on other aesthetic or conceptual elements. In general, the role of visual pleasure in AI-generated art will depend on the specific goals and intentions of the creators of the art.

PL: Does prompt-based AI-based art manipulate the prompt writer to create works that inform the AI system to do more pleasurable works for the prompt writer? Do AI systems train their users as much as the users train the systems?

GPT-3: Prompt-based AI-generated art could manipulate the prompt writer to create works that are more pleasurable to them. Depending on the specific algorithms and techniques used to create the art, the AI system may be more likely to generate works that align with the preferences and biases of the prompt writer.

At the same time, one must note that the prompt writer may also influence the AI system by providing specific input data and instructions that shape the generated art. In this sense, the relationship between the AI system and the prompt writer may be somewhat reciprocal, with both parties influencing each other in creating the art.

The extent that the AI system and prompt writer influence each other depends on the specific algorithms and techniques used to create the art

. These rules also apply to the goals and intentions

of the creators.

PL:

Will traditional artists become obsolete?

GPT-3: It is unlikely that traditional artists will become obsolete due to the development of AI-based art. While AI algorithms can create sophisticated and visually pleasing art, traditional artists bring a unique perspective and skills to the art-making process that a machine cannot replicate.

Traditional artists draw on their experiences, emotions, and creativity to create art reflecting their vision and style. This uniqueness is something that AI algorithms are currently not capable of replicating. As a result, traditional artists will likely play a vital role in the art world, even as AI-generated art

emerges.

It is also important to note that the value of art relates to its uniqueness and authenticity, which may map to works created by traditional artists rather than AI algorithms. As a result, traditional artists may keep their value for their perspectives and original contributions to the art world.

PL: What

1. is the future of AI-based art?
2. GPT-3: It is difficult to predict the future of AI-based art. It will depend on advances in AI and machine learning technologies, art audiences’ evolving tastes and preferences, and AI art’s cultural and social context.
3. One possibility is that AI-generated art will continue to evolve and
4. become more sophisticated, with algorithms capable of creating increasingly complex
5. and nuanced works of art. The increasing sophistication of AI could lead to more applications for AI-generated art, including in fields such as advertising, film, and video games.
6. Another possibility is that AI-generated art will become more integrated into the broader art

world, with AI-generated works exhibited alongside traditional art in galleries and museums. The coexistence of traditional and AI-based art could lead to a greater appreciation and understanding of AI-generated art among the general public.

The future of AI-based art is likely to be complex and dynamic, with many different possibilities and outcomes.