How to Create Generative Art in 2023? [With ChatGPT]

This article was updated on June 10, 2023.

Well-established researchers and historians formed many definitions. This one is one of my favorites so far:

Generative art refers to any art practice where the artist uses a system, such
as a set of natural language rules, a computer program, a machine, or other
procedural invention, which is set into motion with some degree of autonomy
contributing to or resulting in a completed work of art

Philip Galanter,

What is Generative Art? - Complexity Theory as a Context for Art Theory.

We could say code art is a subset of generative art, which, as the name says, uses code to create a piece of art. They usually use mathematical methods and incorporate random functions to ensure each work is unique. This form of art has been explored since programming languages and computers were created, but they have become widely popular with the introduction of the blockchain.

I wrote the difference between generative art and AI art below.

You may well be familiar with the buzz surrounding ChatGPT in recent months (If not, I urge you to give it a go here). The efficiency and user-friendliness of this chatbot have sparked widespread discussion about AI, its present-day applications, and its implications for the future.

You might be wondering, can chatgpt create art? and the answer is yes. Let’s dive in.

Subscribe now

I've seen countless Twitter threads and Instagram videos extolling the virtues of incorporating ChatGPT into one's daily routine to save time. And businesses are adapting to this shift, as evidenced by the emergence of a new job title on job boards - the ChatGPT engineer.

As it turns out, ChatGPT is also rather adept at coding. Whether you need templates or ideas to jumpstart your programming, ChatGPT can quickly generate code in virtually any language.

Although I’ve written about AI art in numerous issues (last week I wrote about an autonomous AI artist), it has been about artists using prompts or training their own generative art models using different techniques to produce visual outputs directly, not code.

Why is this new chatbot so important — in the context of algorithmic art?

Because ChatGPT can create generative art, too.

Kaloh’s Newsletter is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

I’m not the first one to try or point at this; in fact, Pronoia, a popular collector who sold Zancan’s Lushtemples, Highlights of the Hike for 83k Tezos — included in 10 iconic 1/1s on Tezos — was one of the first to explore the concept of AI-assisted generative art back in August 2022 in his article “GPT-3 and me: A generative coding study”.

Here is a snippet of that article (keep in mind, it uses GPT-3, which, although pretty powerful, wasn’t at the same level as GPT-4, the latest version).

I knew GPT could create simple code and source it from a large codebase already presents in the training data. What I wanted to see was whether I could push the system to produce something that could be considered “original”. This article follows selected inputs and outputs from the generative art journey between GPT-3 and me, a collaborative article. I have never written a line of Javascript code in my life.

….

Not only are all the outputs shown in this article produced by GPT through my inputs, the article itself was also partly written by the system. I encourage readers to try and find what parts of this article were purely written by a human, and which GPT wrote. It was extremely fun and rewarding to work with this system, trying to push it’s coding limits. In the future I believe these systems will play an integral part in discovering the potential of creativity in artificial intelligence.

Since then, many more have explored the chatbot power. It isn’t surprising that Claire Silver, who started to explore AI art in 2018, also jumped into AI-assisted coding.

Using GPT4 to write code is quickly *teaching* me code. Its code is 90 percent right, but the 10 percent that isn’t has me learning by doing, and I’m retaining what I’ve learned. Brain feels weirdly neuroplastic in a way it hasn’t since elementary school

Claire Silver

As she highlights, learning by doing has been very useful, as the code isn’t exactly perfect (10% wrong), and you need to fix it.

After seeing these use cases, I wondered, have any AI-assisted long-form generative coded projects been published?

Kaloh @Kaloh_nft

When will we see the first GPT #fxhash collection? Surprised none has done it yet (maybe I am wrong?)

6:51 PM ∙ Mar 22, 2023

52Likes5Retweets

The answer is a handful of projects have been published since mid-2022. The open fxhash platform has been the home for these experimentations, as anyone can release a collection by following their coding guidelines.

Here are 7 generative art code examples that have different levels of artificial intelligence involvement, and they can help you understand how to use chatgpt for art.

Or, should we call these chatgpt art?

The Geometry of Rain by Chris Wallace.

An artwork co-created by Chris Wallace and GPT-4 exploring the relationship between the geometric shapes and the rain, highlighting the intersection of mathematical precision and organic fluidity.

ChatGPT writes, "This artwork explores the relationship between the geometric shapes and the rain, highlighting the intersection of mathematical precision and organic fluidity."

In this case, Chris (founder of UltraDao and the Woodies) used a trial-and-error process to update the prompts and show chatGPT creating art. He recalls that the code needed to be updated manually in some cases. The title was also generated by ChatGPT-4.

Published in July 2022, "Starmap - AI Study I” was created by CoDexter (John Wowkavic & BrainArt Labs) entirely using GPT-3 as only the Math.random() operation was added (using fxrand()) to produce the randomization needed.

Starmap - AI Study I by CoDexter.

Starmap brings you to different sectors of the generative art galaxy. A first of its kind generative art experiment using a Text2Code workflow, which involves writing text prompts that get translated into javascript code. This project was created with English words and human imagination, but the programming of the piece is all AI.

CoDexter continued his exploration of AI-assisted generative art algorithms and teamed up with collector Maxwell White. This collection has a powerful premise “What is the point of coding something so that it looks like it’s not coded?”

Man and the Machine by Maxwell White and CoDexter.

Overall, Man and The Machine is a testament to the creative potential of combining human artistic vision with the power of AI and computer code. Together, we sought to push the boundaries which occur at the intersection between the organic and the inorganic, and this collection is a celebration of the beauty and complexity of both.

Abdi assembled one of the most popular AI-assisted collections until today on fxhash. He used ChatGPT-3 to generate a poem and Dall-E 2 to come up with stunning images. As Abdi describes it, the human touch shows up when merging both technologies and creating the right compositions.

The Beauty of Discord by Abdi.

The combination of these two AI technologies, along with the human touch in the painting algorithms and composition, creates an artwork that is truly one-of-a-kind. The final product is not only visually stunning, but it also serves as a glimpse into the future of humanity and the possibilities of AI collaboration in art. The artwork is a testament to the beauty that can result from a seemingly discordant combination of human and machine inputs, and it invites us to consider the impact of technology on the creative process and the role of the human artist in the age of AI.

Materia is a collection inspired by its resemblance to carved wood, metal, or elements made of silk or felt. In this case, the artist used ChatGPT to generate the function names used as the features, but not in the code — hinting that AI can also assist in creative areas.

Materia by Water Flowing.

Sasha, known for her AI-assisted poetry, has been using AI for a long time. She has published books and multiple collections on fxhash and other platforms together with her AI alter ego, Technelegy. Besides the AI component, this collection is particularly exciting as it was featured on the Here & Now immersive experience and uses fx(params). This new fxhash feature allows collectors to co-create their pieces by adjusting the settings used to develop the final artwork or should we say, chatgpt artwork?

SOUND GROWS INTO SPACE by Sasha Stiles.

Poetry endures across space and the passage of time precisely because of a paradox: it is immutable yet intensely personal, becoming something irrefutably unique with each and every reader. In this poem, shifting parameters influence the flow of feeling and understanding, imbuing verse with multidimensional meaning and empowering deep engagement in the act of reading.

Another collection that uses ChatGPT, specifically to create a poem as the centerpiece of the work. The code isn’t generated by AI, only the poem.

Full Of Flowers by SMLDMS.

Oh my god, it's full of flowers
Nature's art, in full bloom
Petals soft, a colorful array
A sight to behold, they gently sway

But with computers, a new art is born
Generative designs, a digital morn
Code and algorithms, a new way to create
Flowers rendered, no longer innate

Gone are the days of pen and paper
Now, with a click, a new flower caper
Infinite variations, a never-ending bouquet
Generative art, a new way to play

But let us not forget the flowers of old
Nature's beauty, worth more than gold
So let us embrace both, hand and machine
A balance of art, a true scene serene

An interesting example from the past. ChatGPT and future AI iterations will have more profound effects and ramifications than calculators—credits to Ben Tossel for sharing this newspaper snippet.

The implications of AI-assisted art are complex and multifaceted. While some may argue that this new technology threatens to replace human creativity and skill, others see it as an opportunity to explore new artistic horizons and push the boundaries of what is possible.

On the one hand, it offers artists new tools and techniques that can enhance their work in exciting ways (as we saw in the multiple examples). On the other hand, many have been raising questions about authorship and originality. When it comes to generative art, the question is how much you value coding skills.

Purpose, connection, and originality will be even more crucial than technical complexity and the WOW factor.

Until next time,

- Kaloh

PD: If you liked my writing, feel free to click the ❤️ button on this post so more people can discover it on Substack 🙏

Consider subscribing to Kaloh’s Newsletter to receive my articles for free in your inbox. For the full experience, become a premium subscriber.

Subscribe now

• Receive premium and public posts, and access the full archive.

• Access to my private Discord server with over 300 NFT enthusiasts.

• Participate in monthly giveaways.