Connect with us

AI Generative Art

It’s been three years since this technology has rocked every industry. How far has ours come?

mm

Published

on

ABOVE PHOTO: Chat GPT and DallE try to show the evolution of Generative AI artwork over the course of 2022-2024 in this image. It interprets the change as pixelation decrease but in reality the change was more complex and generic. Essentially Generative AI has become a very powerful clip-art tool in this author’s opinion. 

GENERATIVE AI TOOLS for art significantly have changed in the past three years, shifting the creative landscape and sparking debates about ethics and creativity. In 2023, Generative AI art experienced a surge in popularity with tools like MidJourney, Stable Diffusion, and Ideogram leading the charge. These platforms allowed users to create stunning visuals with simple text prompts; users were awed by the experience in the early days. However, the initial “wow” factor began to wane, with many perceiving AI-generated art as generic or artificial.

Midjourney has experienced significant user growth since its launch. Midjourney surpassed 1 million users within six months of its initial release in 2022. By July 2023, Midjourney’s user count had grown to more than 16 million. As of September 2024, Midjourney had 20 million registered users. In 2024 the organizåtion introduced a personalization feature allowing the AI to “get to know” users and potentially improve their creative process. Ideogram also launched Canvas, a tool enabling users to select and replace specific parts of images, giving more control over the generated artwork. Microsoft made a surprising move by integrating AI into its upcoming rerelease of the classic MS Paint application. The update included features like Generative Fill and Generative Erase, powered by Copilot+. With this, users could add or remove elements in their artwork using text prompts.

Also, Recraft emerged as a notable player in the Generative AI art space. This platform offers photorealistic and vector graphic generation capabilities, with an impressive editor independent of its models. While its text accuracy may need improvement, its vector output capabilities are excellent. Also, Canva integrated generative AI technology into its platform, offering a suite of AI-powered tools called Magic Studio. These tools include text-to-image generation, AI-enhanced graphics, and automated design creation. Like in Recraft, Canva’s Background Remover tool lets you delete backgrounds with just one click. Stable Diffusion was a challenge to use in 2022, and today it is available as a Photoshop plugin. Several options are available for integrating Stable Diffusion into Photoshop. Stable.art, a free, open-source plugin, lets you use your local Stable Diffusion installation inside Photoshop. Automatic1111-Photoshop-StableDiffusion-Plugin enables users to run Automatic1111 Stable Diffusion within Photoshop. These plugins allow users to generate AI images and perform inpainting, outpainting, and other Stable Diffusion-based tasks directly within Photoshop. This streamlines the workflow for digital artists and designers. It is much easier than the GPU cloud access needed in early 2023. Finally, Hugging Face offers an easy Stable Diffusion workspace for creating illusions, a valuable tool.



Advertisement
Generative AI in the Screen-Print Industry: How Far Have We Come?

This is the Ideogram interpretation of text in 2022 (left) and 2024 (right). Text generation has come a long way on many of the available platforms. In particular, Ideogram has been leading the way for the past two years.

An Old Friend Adds Something New

Adobe added Generative AI capabilities to Photoshop and Illustrator through Adobe Firefly. One of the easiest platforms for artists to start with, it is integrated into the newest versions of Adobe software. Adobe’s Firefly initiative explores generative AI designed for safe commercial use through licensed content training to ensure ethical and legal compliance. The initiative offers digital creators the opportunities to monetize their skills by participating in “content missions,” contributing to generative AI models that add to Adobe Stock with diverse images. It mainly focuses on filling gaps in training data for more inclusive AI models.

In 2024, Adobe launched the Firefly Contributor Bonus, providing financial incentives to Adobe Stock contributors whose content trained Firefly, considering the total number of approved videos and images contributed over time and the number of licenses generated during a specific period. The Firefly Contributor Bonus average payout reportedly is around $10.

Additionally, Adobe has created specific “Missions” to train Firefly, such as capturing city skylines or famous landmarks through windows or from balconies, with contributors submitting between 500 and 1,000 images per content mission. While Adobe’s approach claims to help creators monetize their work, questions arise about the long-term implications for the creative market and whether this collaboration truly fosters creativity and diversity or potentially exploits creators, leading to fewer and fewer missions as the model evolves.
Concerns about ethical implications, such as plagiarism and copyright infringement, gained prominence in 2024. In the U.S., new legislation like the Transparency and Responsibility for Artificial Intelligence Networks (TRAIN) Act aims to protect copyright holders and enable them to discover if their work was used without permission to train AI models if passed.

To help identify Generative AI content, Adobe added “content credentials” in 2024. Content credentials are digital identifiers attached to digital content. They serve as metadata, offering transparency and verification for various types of digital media including images, videos, and audio files. Content Credentials are available across multiple platforms and contain creator information, creation date, edit history, AI involvement, and cryptographic verification. The challenge is that Content Credentials can be circumvented simply by making a copy of the file, which is part of a normal process of separations or editing a graphic.

Generative AI in the Screen-Print Industry: How Far Have We Come?

In early 2022 Midjourney (bottom left) really struggled to depict “Human Hands.” But by 2024 (bottom right), the images of hands noticeably improved in most artistic styles. 

Finding the Balance

Honest AI disclosure by graphic artists became part of 2024 workflows, with some brands and artists opting out of using AI entirely. Tedeschi Trucks encountered an issue with its fanbase after AI-generated art was used for a concert event at Red Rocks Amphitheater outside of Denver. After the discovery, the band promised not to use AI-generated content in the future and refunded their fanbase. Despite novelty wearing off, Generative AI art remains a valuable tool for personalization and streamlining processes in apparel embellishment. Advanced tools, such as OpenAI’s Sora for video generation, showcase AI’s ongoing progress and potential impact in the creative field. Sora’s recent release raised many questions, even among the eager red team artists that Open AI had hired to test Sora. In an open letter to “DEAR CORPORATE AI OVERLORDS,” the red team testing artists said in part that, “ARTISTS ARE NOT YOUR UNPAID R&D; we are not your free bug testers, PR puppets, training data, or validation tokens.”
As this technology evolves, the industry faces challenges balancing innovation with ethical considerations and copyright protection. Some of the tools to detect AI are being outsmarted easily. Not many are creating tools to protect artists. Opt-out lists or clauses inside of Generative AI platforms are available but their effectiveness remains unclear. Programs like the University of Chicago’s Glaze or Nightshade are meant to protect artists, but adopting these tools rarely is easy or free. Even scanning articles like this for AI-generated content yields different results on different platforms.

For the author writing it or the artist creating it, it is frustrating when you want the world to know that you actually did the work. In 2025, generative AI art tools likely will have more personalization, new editing capabilities, and some sort of licensing models to address copyright concerns. It is less likely (but hopeful) that we will see a better way to identify AI as lines continue to blur and, most importantly, fair compensation for artists.

Advertisement

Advertisement

SPONSORED VIDEO

Let’s Talk About It

Creating a More Diverse and Inclusive Screen Printing Industry

LET’S TALK About It: Part 3 discusses how four screen printers have employed people with disabilities, why you should consider doing the same, the resources that are available, and more. Watch the live webinar, held August 16, moderated by Adrienne Palmer, editor-in-chief, Screen Printing magazine, with panelists Ali Banholzer, Amber Massey, Ryan Moor, and Jed Seifert. The multi-part series is hosted exclusively by ROQ.US and U.N.I.T.E Together. Let’s Talk About It: Part 1 focused on Black, female screen printers and can be watched here; Part 2 focused on the LGBTQ+ community and can be watched here.

Promoted Headlines

Most Popular