“Dear Judith, when I think of your work, AI-generated images immediately come to mind. Would you be comfortable with the label ‘Midjourney evangelist’?”

I find that term rather amusing. While I’ve worked with Midjourney for over two years and consider it the superior tool for my purposes, I wouldn’t defend it unconditionally. If a better solution emerges, I’m quite prepared to switch allegiances. The competition is making remarkable progress.

Which other image generation tools have caught your attention recently?

Ideogram, for instance, excels at integrating images and text. For placing slogans on marketing materials, it proves far more reliable than Midjourney, which still struggles with typography. However, we’re witnessing intense competition among AI image generators, leading these tools to become increasingly similar in their capabilities.

So eventually, will they all offer comparable features, making the choice between them merely a matter of preference, like Apple versus Microsoft?

It’s not quite that straightforward. One crucial distinction lies in the quality of the training data, where Midjourney holds a significant advantage. When launching the company in early 2022, CEO David Holz specifically sought out artists, photographers, illustrators, and architects. In contrast, DALL-E trained primarily on material from Wikipedia. This fundamental difference remains evident in their outputs today.

How does this difference in quality manifest itself?

It’s apparent in the attention to detail—particularly in rendering human eyes. It also shows in the handling of focus: Midjourney allows precise control over where the lens concentrates—whether on the foreground, background, or across the entire image. Overall, its AI demonstrates superior management of mood and lighting, adhering to higher aesthetic standards.

Where have you found Midjourney particularly effective?

It excels in social media applications. The tool enables rapid creation of customized images. Once you develop a visual style that aligns with your corporate design and identity, you can deploy it consistently to strengthen your brand presence. Midjourney offers newcomers a wide range of predefined styles, while experienced users typically develop their own visual language.

What are the current limitations of AI image generation?

They become apparent in specialized fields such as medicine, botany, or biology. Recently, a client requested anatomically accurate representations of muscle structures in horses and humans. This proves impossible at present—and likely will remain so for some time—because the AI lacks training in medical imagery. Even accurately depicting specific dog breeds poses challenges. For expert applications, you must either forgo AI images entirely or have them verified by specialists. There’s no middle ground.

How do you navigate copyright issues?

The legal landscape remains uncertain. While prompt creators clearly own their inputs, image ownership proves more complex. Legal experts tend to view AI-generated images created through simple prompting as non-copyrightable. However, images involving significant human creative input—whether through post-processing or sophisticated prompting—may qualify for copyright protection.

The boundaries remain decidedly unclear. Midjourney grants users copyright while retaining an unlimited and irrevocable right of use, meaning the platform can also utilize my images. This poses challenges for companies requiring exclusive content. It’s essential to scrutinize each AI platform’s terms of use carefully. Midjourney demonstrates more transparency than its competitors in this regard.

The monthly subscription of $8–10 seems modest. But what about the time investment? What should one expect to spend for a high-quality image?

It varies considerably. Sometimes I can achieve the perfect image in five minutes. Other times, I might work for two hours and still feel dissatisfied. But doesn’t searching traditional photo databases or commissioning photographers also consume time? Ultimately, AI proves most cost-effective for my needs. However, the real value lies elsewhere: this technology offers a new avenue for expressing creativity.

Could you elaborate on that?

Recently, I created a poster for a concert featuring Bizet’s Carmen and Gershwin’s Porgy and Bess. The atmospheric images generated through Midjourney resonated strongly with the client. The project proved particularly satisfying because I could implement my own creative vision rather than rely on stock images. Though already familiar with both operas, I delved deeper into their themes, motifs, and symbols.

This comprehensive understanding proved crucial for crafting effective prompts and, consequently, generating compelling images. In essence, AI amplifies existing strengths, enabling more direct and powerful expression of creative ideas.

Interview: Verena Parzer-Epp