A landmark study of nearly 400 professional visual artists reveals an overwhelming rejection of generative AI tools like Midjourney and DALL-E, with 99% of respondents expressing strong dislike for the technology. Researchers at Carnegie Mellon University surveyed working artists about how AI image generation has changed their livelihoods, opportunities, and outlook, presenting their findings at the Association for Computing Machinery (ACM) CHI conference in April. The results paint a stark picture of an industry in crisis.

What's Really Happening to Artists' Livelihoods?

The study's most striking finding is the sheer scale of artist rejection toward generative AI. According to the research, 92% of artists who dislike AI do so "strongly," and this distaste translates directly into action. An overwhelming 85% of artists completely abstain from using AI tools at all, while 88% refuse to use AI specifically for image generation. This represents one of the most unified rejections of any technology across an entire profession.

Yet despite this widespread refusal, artists encounter AI-generated images constantly in their work. The study found that a majority of artists encounter AI-generated images at work every week, and many see them daily. This happens because clients and employers have begun sending artists AI-generated images for "brainstorming" or for direct touchup work. The result is a profession experiencing what researchers describe as "the degradation of a skilled trade in realtime."

Most artists report feeling they are now competing directly with AI-generated artwork on the marketplace. Well over half say that opportunities and pay rates are declining as clients and employers embrace generative AI. Artists are increasingly offered touch-up jobs rather than well-paying, sustainable creative gigs. The emotional toll is evident: pessimism, anger, and even despair have become common sentiments across the field.

Why Are Artists So Opposed to These Tools?

Artist concerns about generative AI tools like Midjourney stem from both ethical and practical objections. The research identifies two key categories of resistance:

• Ethical Concerns: Many artists learned that tools like Midjourney and NightCafe were sourcing training data from artwork without the permission of original creators, leading them to stop using the tools entirely
• Functionality Issues: Artists report that AI image generators consistently fail to produce useable results, with one illustrator and designer noting that tools "have consistently failed to produce useable results" despite experimentation
• Intrusive Technology: Researchers concluded that "participants viewed genAI as an intrusive technology within the creative domain," fundamentally incompatible with how artists see their craft

"I experimented with NightCafe and Midjourney when they were being developed (in beta). When I learned they were sourcing their art without permission of the artist, I stopped using them," one survey participant stated.

Professional Visual Artist, Carnegie Mellon Study

The research suggests that while AI image generation hasn't completely replaced art in many cases, likely because the output is too homogenous and unappealing to the public, the technology is grinding down the craft itself. It's robbing visual arts of key human elements and pushing down wages, pay rates, and opportunities across the board.

How Are Detection Systems Catching AI-Generated Images?

As AI image generators like Midjourney become more sophisticated, researchers are developing better ways to identify synthetic images. A separate study introduced AIText2Image, a large-scale dataset of photorealistic AI-generated images from modern text-to-image generators, and trained detection models to identify fake images with robust accuracy.

The research integrated 16 different explainable AI (XAI) methods into a detection framework to help humans understand why an image is classified as AI-generated. The study found that humans can correctly classify only 61% of images as real or synthetic, while state-of-the-art AI detectors largely surpass this performance. The detection models were trained on images from multiple generators, including Midjourney 5.2, SDXL 0.9 and 1.0, DALL-E 2 and 3, Adobe Firefly, and others.

This work is critical as generative AI becomes more accessible. The misuse of AI-generated images in online disinformation campaigns, deepfakes, and fake news poses significant societal risks. Governments worldwide, including the European Union with its AI Act, are developing regulations to mitigate the negative impact of synthetic media.

Steps Artists Can Take to Protect Their Work and Careers

• Document Your Process: Keep detailed records of your original work, including sketches, drafts, and timestamps, to establish clear ownership and originality in case of disputes over training data usage
• Understand Your Rights: Join professional organizations and legal initiatives that are pursuing class action lawsuits against AI companies for unauthorized use of artist work in training datasets
• Communicate Boundaries: Clearly communicate to clients and employers that you do not use AI-generated images in your workflow and explain the ethical and quality concerns that drive this decision
• Develop Specialized Skills: Focus on areas where human creativity and judgment are irreplaceable, such as conceptual work, character development, and narrative-driven visual storytelling that AI tools struggle to execute

What Does This Mean for the Future of Creative Work?

The Carnegie Mellon study offers a crucial insight: the widespread rejection of AI tools by artists themselves suggests that generative AI may not be the wholesale replacement of creative labor that some predicted. However, the technology is still causing real economic harm through wage suppression and reduced opportunities.

The research also highlights a fundamental mismatch between what AI image generators can do and what creative professionals actually need. While tools like Midjourney can produce visually interesting images quickly, they lack the intentionality, originality, and problem-solving capability that professional artists bring to their work. The homogenization of AI-generated content, combined with the ethical concerns about training data, has created a situation where 99% of professional artists actively reject the technology.

As detection systems improve and regulatory frameworks develop, the landscape for AI image generation may shift. But for now, the message from working artists is clear: generative AI is not a tool they want, and it's not solving problems they actually face. Instead, it's creating new ones.