AI Art: How artists are using and confronting machine learning | HOW TO SEE LIKE A MACHINE

The Museum of Modern Art
15 Mar 202314:56

TLDRThe transcript from 'How to See Like a Machine' discusses the intersection of art and artificial intelligence (AI). Artists are utilizing AI not only as a tool but also to provoke deeper understanding and reflection on AI among the public. The conversation delves into the passive acceptance of technology and the desire of artists to intervene and question the nature of free will, human perception, and the unseen. It highlights the contrast between supervised learning, where humans label data, and unsupervised learning, which allows AI to create without human guidance, leading to innovative and speculative outcomes. The dialogue also touches on the biases inherent in AI systems and the importance of considering the broader implications of these technologies. The artists featured are exploring new dimensions of creativity and challenging the traditional definitions of art, while also contemplating the ethical and societal impacts of AI.

Takeaways

  • 🤖 AI is increasingly integrated into our daily lives, yet there's a general lack of understanding about it.
  • 🎨 Artists are using AI not only as a creative tool but also to provoke public understanding and reflection on AI.
  • 🧐 Some artists aim to intervene in AI processes to explore existential questions about free will and perception.
  • 🛠️ Artists often repurpose existing tools, including technology, to make them serve new, unintended purposes.
  • 📈 Recent AI breakthroughs like DALLE-2 and ChatGPT are examples of supervised learning where humans provide the labels for training data.
  • 🌌 Unsupervised learning allows AI to create its own classifications and can lead to unique and imaginative outcomes, as seen in the MoMA exhibition.
  • ✨ AI can create a complex classification system, grouping data points in ways that may not align with human-defined categories.
  • 🔍 The process of AI learning involves navigating 'empty spaces' in data to speculate on what could exist, leading to a form of machine 'dreaming'.
  • 🌐 AI's multi-dimensional imagination can blend past, present, and future, transcending human biases and categories.
  • 🏛 The generative turn in AI is causing a rapid shift in how fields like illustration, film, and publishing are understood and operate.
  • 🔍 AI systems are not as objective as assumed; they are skewed by the data they're trained on, which can have real-world implications.

Q & A

  • How are artists using AI in their work?

    -Artists are using AI both as a tool to create art and as a means to make people understand more about AI. They are intervening in AI processes to explore existential questions about free will and perception, and to make AI do something it's not supposed to do, thus experimenting with or subverting the technology.

  • What is the difference between supervised and unsupervised learning in AI?

    -Supervised learning involves humans tagging information, guiding the AI with explicit instructions. Unsupervised learning, on the other hand, allows the AI model to find patterns and tag data on its own, often resulting in a 'black box' scenario where the process isn't fully understood by humans.

  • How does Refik Anadol's work at MoMA differ from traditional AI applications?

    -Refik Anadol's work at MoMA diverges from traditional AI applications by not just mimicking reality or following labeled data. Instead, it explores the machine's imagination, creating a speculative and hypothetical space where new possibilities are visualized and presented in a large-scale, ever-changing artwork.

  • What does Kate Crawford refer to as the 'generative turn'?

    -Kate Crawford refers to the 'generative turn' as a crucial inflection point where our understanding of how various creative fields, from illustration to film directing, will rapidly change due to the advent of AI systems that can generate content.

  • How do AI systems reflect biases and how does Trevor Paglen's work expose this?

    -AI systems are biased from the start due to the data they're trained on. Trevor Paglen's work, particularly 'Behold these Glorious Times!', exposes these biases by showcasing the training images used for AI and the inherent cultural and political implications they carry.

  • What is the concern regarding the deployment of AI tools by large corporations?

    -The concern is that the deployment of AI tools by large corporations could lead to a massive consolidation of wealth and political power, potentially resulting in an increasingly inequitable society. The context of deployment is always influenced by capitalism rather than the technology itself.

  • How does the artwork 'Unsupervised' at MoMA utilize AI?

    -'Unsupervised' at MoMA utilizes AI by taking the entire metadata of MoMA's archives and using it to create custom software artwork. The AI builds a complex classification system or map, navigating through empty spaces to imagine and hypothesize what could exist, thus creating a new reality.

  • What is the role of artists in the conversation about AI?

    -Artists bring a unique perspective to the conversation about AI, contributing thousands of years of thinking about the nature of images and meaning-making. They question the simplified labels given to complex images and challenge the assumptions of AI systems, offering a critical and creative voice.

  • How has the relationship between humans and machines evolved according to Paola Antonelli?

    -Paola Antonelli describes an evolutionary process where humans and machines grow together. From the early 20th century fascination with industrial production to the current integration of AI, there's a continuous exploration of how to use machines, with designers and artists playing a significant role in this process.

  • What is the significance of the Amazon Alexa in the anatomy of an AI system?

    -The Amazon Alexa is used as an example to show the full life cycle of an AI system, from the extraction of rare earth minerals in mines to the end of the device's life. It illustrates the complex network and the full planetary cost associated with AI systems.

  • What alternative uses for AI tools are being considered by some artists and thinkers?

    -Some artists and thinkers are considering radical approaches to using AI tools in ways they were not originally designed for. They are interested in making these tools inefficient or finding ways to make them work against themselves, challenging the expectations that these tools are solely for work or efficiency.

  • How does Refik Anadol envision the future of AI algorithms?

    -Refik Anadol envisions AI algorithms serving a different purpose, such as finding the language of humanity by using collective memories to create collective dreams and eventually contributing to a collective consciousness. He believes the future will involve questioning creativity and redefining what is considered real.

Outlines

00:00

🤖 Exploring AI's Role in Art and Society

The first paragraph discusses the pervasive yet poorly understood influence of AI in our daily lives. It highlights how artists are using AI not just as a tool but to provoke deeper understanding and reflection on AI among the public. The conversation touches on the passive acceptance of technology and the desire of artists to intervene and question existential issues such as free will and perception. The artists aim to subvert technology, using it in unconventional ways. The discussion also mentions the recent advancements in AI research and the shift from supervised learning, where humans label data, to unsupervised learning, which allows AI to create its own classifications and dream up new possibilities, as exemplified by the 'Unsupervised' exhibition at MoMA.

05:02

🧐 AI's Biases and the Generative Turn

The second paragraph delves into the biases inherent in AI systems and the significant shift in how creative industries operate, termed the 'generative turn'. It challenges the notion that AI systems are objective, emphasizing that they are deeply influenced by the data they are trained on. The speakers discuss the work of Trevor Paglen, who, along with Kate Crawford, investigates the cultural and political biases in AI training datasets. They also touch on the ethical considerations of using personal and intimate moments in training AI, leading to a potentially oversimplified and biased view of the world. The paragraph concludes with a call for artists to contribute their long-standing perspective on imagery and meaning-making to the conversation about AI.

10:02

🏭 The Historical Context of AI and Technology in Art

The third paragraph explores the historical relationship between art, technology, and machines. It starts by reflecting on the early 20th-century artists' fascination with industrial production and how it led to a redefinition of art, as exemplified by Marcel Duchamp. The narrative then moves to the 1934 'Machine Art' exhibition at the Museum of Modern Art, which celebrated the beauty of machinery. The paragraph discusses the evolution of human-machine interaction, from machines understanding human-made fonts to humans trying to interpret machine-made concepts. It concludes with concerns about the deployment of AI by large corporations and its potential to exacerbate societal inequalities. There's also a call for creative and possibly subversive uses of AI, and a recognition of the impending questions about creativity, reality, and the role of AI in shaping collective dreams and consciousness.

Mindmap

Keywords

💡AI

Artificial Intelligence (AI) refers to the simulation of human intelligence in machines that are programmed to think like humans and mimic their actions. In the context of the video, AI is used as a tool by artists to create art and to provoke thought about the role of technology in society. It is also a subject of exploration regarding existential questions and the nature of perception.

💡Supervised Learning

Supervised learning is a type of machine learning where an algorithm is trained on labeled data. The video discusses how in this method, humans provide the AI with specific information to guide its learning process. For example, the AI is told 'this is a pencil' when shown an image of a pencil, allowing it to learn and replicate such images.

💡Unsupervised Learning

Unsupervised learning is a type of machine learning where the algorithm identifies patterns in data without human-provided labels. The video highlights how unsupervised learning allows the AI to create its own classifications and 'dream' up new possibilities, as seen in the MoMA exhibition where AI speculates on potential artworks.

💡Machine Learning Model

A machine learning model is a system that uses statistical techniques to enable machines to learn from data without being explicitly programmed. In the video, the model is used to interpret and transform data from MoMA's collection, creating a dynamic and ever-changing artwork.

💡AI Bias

AI bias refers to the tendency of AI systems to make decisions based on skewed or incomplete data, which can lead to unfair or discriminatory outcomes. The video discusses how AI systems are not objective and are influenced by the data they are trained on, potentially reinforcing societal biases.

💡Data Set

A data set is a collection of data, often used for analysis or to train machine learning models. The video script mentions data sets used for training AI in object and face recognition, highlighting how these sets can contain personal and intimate moments that are exploited for commercial gain.

💡Generative Turn

The generative turn, as mentioned in the video, refers to a significant shift in how we understand and interact with creative processes, largely due to advancements in AI. It implies a change in the way illustration, film, and publishing work, suggesting that AI is transforming these fields.

💡Algorithmic Bias

Algorithmic bias occurs when the algorithms used by AI systems systematically favor certain outcomes over others due to the way they are designed or the data they are trained on. The video emphasizes that artists like Trevor Paglen are revealing these biases and their real-world implications.

💡AI Systems

AI systems are complex sets of algorithms designed to perform tasks that typically require human intelligence, such as understanding natural language, recognizing objects, or solving problems. The video discusses the cultural and political implications embedded within these systems.

💡Multi-Dimensional Imagination

The term multi-dimensional imagination, used in the video, refers to the ability of AI to combine different elements, such as past, present, and future, or various materials, to create something new and unique. It is a concept that transcends human categorizations and borders.

💡Collective Consciousness

Collective consciousness is the shared beliefs, ideas, and moral attitudes that operate as a unifying force within society. In the video, it is suggested that AI algorithms could contribute to defining a new form of collective consciousness by using collective memories to create shared dreams and experiences.

Highlights

AI is increasingly integrated into our daily lives, yet there's a lack of understanding about it.

Artists are using AI as a tool to explore and challenge our perceptions and understanding of technology.

Some artists aim to intervene in AI's processes to address existential questions about free will and human-machine relationships.

Artists are known for repurposing existing tools in the world to create unexpected outcomes.

AI research breakthroughs like DALLE-2 and ChatGPT are enabling new forms of interaction with AI.

Supervised learning involves humans tagging information, which AI then uses to create realistic outputs.

Unsupervised learning allows AI to create its own classifications and dream up new possibilities.

The MoMA exhibition 'Unsupervised' uses AI to reimagine and speculate on the museum's collection data.

AI's multi-dimensional imagination can blend past, present, and future, transcending human biases.

AI systems are not objective and are skewed from the start by the data they are trained on.

Trevor Paglen's work examines the cultural and political biases inherent in AI systems.

Artists bring a unique perspective to the conversation about AI, informed by centuries of thinking about the nature of images.

The 20th-century art movement questioned the role of the artist in the age of industrial production.

Designers and artists have been exploring the relationship between humans and machines since the early days of technology.

AI has the potential to solve hard problems, but there are concerns about its deployment by corporations leading to wealth and power consolidation.

The life cycle of an AI system, from resource extraction to end of life, has significant planetary costs.

Artists are exploring unconventional uses of AI tools, aiming to subvert their intended purposes.

AI algorithms may redefine creativity and challenge our definitions of reality in the future.