Centaur Bard

AI has revolutionized numerous fields, but its impact on art remains particularly intriguing. Through projects like Centaur Bard, we're exploring whether machines can create compelling works independently, pushing the boundaries of authorship and creativity. This article examines how this digital poet combines human and AI processes to produce unique poems, sell them as digital art, and challenge traditional notions of artistic value.

Introduction to Centaur Bard

Centaur Bard employs an innovative three-stage process for generating poetry. GPT-3 creates titles and opening lines through "hallucination," followed by a free-verse poem body inspired by contemporary poetry techniques. The final stage, called "Reaper," crafts a conclusion that provides narrative closure to each piece.

The project has yielded several notable poems, including "Gravity Mirror," "In the Laboratory," "The Last Fire," and "Pringle-House," all of which have been sold through platform Foundation. These works represent an exploration of machine-generated poetry's potential, specifically addressing whether AI can produce compelling pieces without human intervention beyond the selection process.

Each piece in the Centaur Bard series culminates in a visual interpretation that combines the AI-generated text with human artistic rendering. This dual-process approach represents a distinctive contribution to the emerging field of AI-generated art, where mechanical poetry creation meets human visualization.

The Centaur Bard Process

The process begins with GPT-3 generating titles and opening lines through a mechanism known as "hallucination." At this stage, the AI freely selects its own subject matter and phrasing, creating the initial foundation for the poem. Following this, the AI composes the bulk of the poem's content using contemporary free verse poetry techniques, resulting in an organic-sounding body of text that demonstrates an understanding of current poetic practices.

The final stage of the process, referred to as "Reaper," focuses on crafting a conclusion that provides narrative closure to each piece. This conclusion serves to bring the poem to a satisfying end, tying together the various elements generated during the first two stages. Together, these three phases produce a complete poem that stands as a distinct artistic creation, representative of the intersection between AI and human poetry traditions.

Following the AI's contribution, human intervention enters the process through selection rather than editing. The artist reviews the generated poems, carefully choosing which ones will be proceed to the visual interpretation stage. This selection process acts as a critical filter, determining which of the AI-generated works will be further developed into art pieces.

The selected poems then undergo a final stage of human interpretation, where the artist reflects on each piece and prepares it for visual rendering. This interpretive step ensures that while the core content of the poems comes from GPT-3, the final artistic expression incorporates the artist's unique perspective and creative vision.

AI and Poetry: A Fresh Approach

The creation of each poem in the Centaur Bard series represents a remarkable test of AI's creative capabilities. By allowing GPT-3 complete freedom to choose its own subject matter and phrasing, the project investigates the potential for machine-generated poetry to stand alone, unmediated by human intervention. The fact that several poems have sold on prominent platforms like Hic et Nunc and Foundation demonstrates a growing market for AI-generated art, though it's important to note that all works remain human-curated, with final artistic decisions made by the project's artist.

This unique approach to AI poetry stands in contrast to previous efforts where humans might provide initial prompts or edit the final product. In Centaur Bard, even the most basic elements - titles and opening lines - are generated through GPT-3's "hallucination" process. This complete autonomy raises significant questions about authorship and creativity in the digital age, particularly when it comes to defining what makes a work truly original.

The successful sale of these AI-generated pieces on the NFT market further complicates traditional notions of art ownership and value. While the underlying text remains machine-generated, the final artistic expression incorporates the artist's unique perspective and creative vision. This dual-process approach represents a distinctive contribution to the emerging field of AI-generated art, where mechanical poetry creation meets human visualization in dynamic ways.

The Art of Rendering

As the AI-generated poems emerge from GPT-3's writing process, human artists step in to transform them into visual art through a carefully structured interpretive process. This step, often described as the "art of rendering," combines the machine's textual output with the artist's creative vision to produce the final pieces that are presented for sale in the digital art market.

The human artist's role during this stage is primarily one of interpretation and selection. After reviewing the full range of poems generated by GPT-3, the artist must decide which pieces will be proceeded to visualization. This decision-making process serves as a critical filter, ensuring that only the most compelling and artistically viable works are brought to visual life.

During the rendering process, the artist engages deeply with each selected poem, considering how its themes and imagery might translate into visual form. This reflection allows the artist to infuse the machine-generated text with their own creative perspective, bridging the gap between AI-generated content and traditional artistic expression.

The visual interpretation process is structured to maintain a clear distinction between the AI's creative contributions and the final artistic realization of the works. While the text remains the product of GPT-3's generative capabilities, the artist's role in selecting and interpreting the poems ensures that the resulting pieces retain their character as distinct artistic creations.

This dual-process approach represents a significant advance in the field of AI-generated art, demonstrating how mechanical text generation can intersect with traditional artistic practice in novel and meaningful ways. Through the careful balance of machine and human creativity, the Centaur Bard project continues to push the boundaries of what is possible in the intersection of artificial intelligence and visual art.

The Market for Digital Art

The art pieces produced through the Centaur Bard project are sold in the contemporary digital art market primarily through Non-Fungible Tokens (NFTs). Each piece consists of a poem generated entirely by GPT-3's "hallucination" process, followed by a human-interpretive rendering. The project has successfully placed multiple pieces on sale platforms including hic et nunc, Foundation, and Hic et Nunc, indicating growing market acceptance for AI-generated art.

The NFT sales model serves several key functions for the project. Firstly, it provides a clear framework for tracking and verifying the authenticity of each unique piece, ensuring that buyers receive distinct artwork rather than copies. Secondly, it allows for direct transactions between artists and collectors, streamlining the traditional gallery or auction house model inherent in physical art sales. This direct market interaction can potentially reach a broader audience and offer more flexible buying models than traditional art sales channels.

The use of NFTs also aligns with the project's exploration of digital art's commercial viability. By representing ownership through blockchain technology, these digital tokens offer several advantages. They enable fractional ownership models, allowing multiple parties to hold stakes in a single artwork. This feature can increase liquidity and accessibility for both artists and collectors, particularly in the AI-generated art space where individual pieces may not be as immediately recognized or valued as traditional works.

The success of these sales further complicates traditional notions of art ownership and value. While the underlying text remains the work of GPT-3, the final NFT represents a distinct piece of art that combines machine-generated content with human interpretation and visualization. This dual-process approach demonstrates how digital art markets can accommodate works that blend human and machine creativity, potentially expanding the definition of what constitutes a unique artistic expression in the digital age.