DVT 25 Years of Service


At DVT, we run regular online events focused on the latest technology trends within the IT industry and invite guest speakers to share their knowledge and insights on various topics. The DVT Insights Events aim to enlighten you, educate you and often, provide a new view on a burning issue within the technology space.

An evaluation of AI design tools: Midjourney, Photoshop AI, Galileo
Jamie Chisholm
Lead UX Designer at DVT

An evaluation of AI design tools: Midjourney, Photoshop AI, Galileo

Wednesday, 28 February 2024 12:03

Imagine a world where tedious design tasks melt away, replaced by a digital assistant whispering possibilities. Artificial intelligence (AI) is rapidly painting this future, offering designers a vibrant new palette to boost efficiency and ignite fresh perspectives.

Yet, with every powerful tool come drawbacks. As seasoned designers, even with decades of experience, it's natural to feel a flicker of uncertainty when faced with these ever-evolving technologies. The sheer speed and automation these tools provide can raise questions about their potential impact on our roles.

A Brushstroke of Efficiency

Gone are the days of pixel-perfecting mock-ups one by one. AI tools like DreamStudio or Midjourney paint entire landscapes with a single text prompt, empowering designers to explore more concepts faster. Meanwhile, ClickUp orchestrates design workflows with clockwork precision, freeing minds for strategic thinking and deeper exploration.

These are just a few strokes on the ever-expanding canvas of AI tools available:

  • Adobe Sensei - Image Generation
  • Canva - Text generation, Image editing
  • UIzard/Galileo - UI Generation

...and more emerging every day. However, it's important to remember that AI isn't a magic wand. Patience is key when wielding these tools, as results might not always perfectly align with your vision.

Even with familiar tools like Photoshop, the built-in AI engine, while powerful, can be tricky to master. Sometimes, settling for the initial output seems easier than spending hours coaxing it to understand your exact vision. This is especially true for those who struggle to articulate their vision clearly.

The Pros and Cons of the AI Artist

To explore the potential and limitations of AI tools, I experimented with three popular options: Photoshop’s AI engine, Galileo, and Midjourney. Below are the context, observations and learnings from the process.

1. Photoshop AI Evaluation Update
  • Task Context: The goal was to replace an image of a house with castle image which is part of Photoshop’s initial intro into its AI engine.
  • Actual Prompt: The prompt given to Photoshop AI was, "Generate a large castle."
  • Outcomes and Visuals: Photoshop AI quickly generated a castle that matched the prior buildings dimensions. However, adjusting the size of the Castle itself proved challenging, highlighting the tool's limitations without proper guidance and knowledge.
  • Improvements: The initial prompt correctly created the requested object, but specifying dimensions or adding keywords related to the mood or lighting could refine the results further. For instance, "Generate a mystical castle, 1920x1080, in a lush, enchanted forest at sunset."

  • Learnings: This experience underscores the importance of precise prompts and an understanding of the tool's scope. It also suggests combining AI-generated content with manual adjustments or other software for final touches.
2. Galileo Evaluation Update
  • Task Context: The objective was to produce initial designs for a mobile app to provide parents with an overview of their children’s school journey.
  • Actual Prompt: "Design five mobile app screens for a school allowing parents to view their children’s reports and progress, as well as communicate with the teachers."
  • Outcomes and Visuals: Galileo produced five designs within seconds, offering a starting point for the project. However, the designs lacked uniqueness and depth, necessitating further customisation to achieve a truly engaging user experience.
  • Improvements: To enhance originality, the prompt could be adjusted to include more specific design elements, such as "Incorporate colours and minimalist design principles for a simple user interface, focusing on usability."

  • Learnings: Galileo's strength in rapid prototyping is evident, but its output needs refinement. Designers must be ready to iterate on these initial ideas, emphasising the tool's role in speeding up the creative process rather than completing it. The images are also completely static and non-editable, necessitating more software to continue.
3. Midjourney Evaluation Update
  • Task Context: The task was to generate a new logo for our company ‘DVT’.
  • Actual Prompt: "Create a fresh logo for a software development company called DVT."
  • Outcomes and Visuals: Midjourney produced four visually striking, if generic, concepts but failed to closely follow the specifics of the brand's identity, product focus and even the basics of what defines a “logo”. The designs were not particularly creative and far too detailed, requiring significant adjustments to align with the brand's vision.
  • Improvements: A more detailed prompt might include specific brand elements, colours, and the type of energy the brand wishes to convey, such as "Use vibrant colours and dynamic shapes to present a software development brand that emphasises a ‘people first’ view."

  • Learnings: This underscores the potential of AI in generating creative inspiration and highlights the need for precise, detailed prompts and subsequent designer input to achieve specific marketing and branding objectives.

Overall, these experiences reveal that AI tools offer exciting possibilities for design, but they are not replacements for human creativity and judgement.

Beyond Automation: The Human Touch Endures

It is undeniable that AI automates tasks with impressive speed, but it shouldn't be seen as a replacement for human expertise and judgement. Rather, it offers exciting possibilities to augment our skill sets, explore more ideas faster, and focus on the aspects that require our unique judgment and creativity.

However, this vibrant canvas also holds drawbacks. Like any powerful tool, AI demands mindful consideration. Subscription costs, server dependence, and ethical concerns require careful navigation, as they will invariably impact your client's budget if left unchecked.

“Remember, AI should be seen as a valuable supplement, not a replacement for a solid design team.”

While AI offers the ability to automate the mundane, it can't replicate the human touch. AI-generated images often skirt the edge of the uncanny valley, appearing slightly off to the human eye. It also lacks the empathy to understand a small business owner's aspirations or a large corporation's need for "fresh, out-of-the-box thinking."

Collaboration: The True Masterpiece

The true magic lies in collaboration - designers wielding AI tools to amplify their vision, not replace it. Imagine crafting mock-ups in minutes, iterating on colours based on real-time user response data, or mining design trends in seconds. These are the strokes that redefine efficiency, not simply automate it.

While AI helps designers paint faster, will the next generation learn to truly hold the brush themselves? It's crucial to ensure they develop both design fundamentals and the ability to use AI tools effectively. Much like AI itself, the progression of design will very much hinge on human input, forging design changes for the next generation.

In conclusion, AI isn't a replacement but a partner in exploration. By embracing its strengths and acknowledging its drawbacks, designers unlock a future where efficiency fuels creativity, and every brushstroke leads to a more impactful design landscape. Even the most advanced AI tools lack the human touch – empathy, critical thinking, and the ability to translate intangible concepts into effective designs. Our ability to understand user needs, tell stories, comprehend the human psyche and inject meaning into our work will always be essential.

About the author

Jamie Chisholm has been a dedicated creative professional in design for nearly 30 years, active since the late '90s. His skills cover freelance, print, digital, UI, UX, and mobile design, and he has worked on projects for well-known companies like Discovery, Multichoice, Telkom, Unilever, Apple, and Disney. Jamie has over 16 years of experience in web and mobile development, where he’s known for creating engaging interfaces and effective workflows.

With a solid history of success and a passion for creativity and technology, Jamie also enjoys teaching new designers through lectures and his written work.

Connect with Jamie on LinkedIn.

DVT 25 Years of Service