The artistry of AI

Designflow

Design has always evolved alongside the tools we use. From early handcrafted methods to digital design systems, each shift has been driven by new ideas, techniques, and technologies that reshape how we create.

With all the available AI models, we’re entering another major turning point. Instead of relying solely on manual execution, designers can integrate AI into their workflow to speed up ideation, streamline iteration, and accelerate production. 

Modern design workflows increasingly use AI to bridge the gap between rough concepts and final assets, reducing time to market. But this transformation isn’t without its trade-offs.

In this article, we’ll explore how AI is reshaping the design world, both the opportunities it unlocks and the challenges it brings.

Wake up, you’re dreaming!

To get a grasp of where the ball really started rolling, we need to take a step back to the year 2015, where google engineer Alexander Mordvintsev went and created a computer vision program called DeepDream. This was a foundational moment in the history of AI generated imagery. DeepDream is a form of ‘inceptionism’ and aims to visualize what a neural network “sees” by enhancing features that are already present in an image. It in turn popularized the ability of neural networks to generate art and revealed how models interpret visual data, paving the way for later techniques like Neural Style Transfer and Generative Adversarial Networks (GANs).

• NST:

Aims to blend the content of one image with the style of another

• GANs:

Are a framework consisting of two competing networks (generator and discriminator) that create new, realistic data from scratch. 

After Google published their techniques and made it open source, several tools like web services, mobile apps and desktop software appeared on the market to enable users to transform their own photos. 

Throughout the years, tools like Runway ML made machine learning more accessible to artists and designers without requiring deep technical knowledge. Around 2021, approaches like VQGAN+CLIP allowed people to generate images from text prompts in a way that felt both creative and controllable, marking an important turning point in generative image creation.

Shortly after, OpenAI introduced DALL-E, demonstrating that AI could reliably translate language into detailed visuals. In 2022, the beta release of Midjourney brought these capabilities to a much wider audience. By allowing users to generate images through simple Discord commands or web interface, it made AI image creation more accessible and intuitive than ever before.

Where we stand with the models

Today, there are many tools available that help designers create modern interfaces and speed up their workflow. Software like Figma, Adobe InDesign, Adobe Photoshop, Adobe Illustrator, and CorelDRAW are just a few examples, and the list continues to grow. As competition increases, companies are constantly trying to improve their tools, especially when it comes to UI and UX design. One of the biggest shifts in recent years is the growing role of AI within these platforms.

A strong example of this is Figma’s newer features, such as “Figma Make,” which allow designers to generate layouts and interfaces in a matter of seconds. By simply describing a design or website in text, the tool can produce a visual starting point almost instantly. While the results are not always perfect on the first attempt, they provide a solid foundation that designers can refine and build upon. In some cases, these tools can even generate code, giving developers a clearer direction for implementation. Similarly, companies like Anthropic are exploring design-focused capabilities in models like Claude, further expanding what AI can contribute to the creative process.

As these technologies continue to evolve, they are becoming more capable of producing realistic and high-quality design explorations. This not only speeds up the early stages of design but also changes how designers approach ideation and problem-solving. While human input remains essential, AI is increasingly becoming a collaborative partner, helping push creative boundaries and making the design process more efficient than ever before.  

Creativity with AI

The emergence of AI has added a new dimension to the creative process, allowing designers and artists to explore ideas more quickly and push beyond familiar boundaries. It can enhance creativity in several ways. For example, AI can analyze large amounts of data to generate concepts that spark new ideas. It can also support visual exploration, where generated images act as starting points for further design. Beyond visual design, AI is now used in areas like music composition, helping create melodies or harmonies, and in writing, where prompts can help overcome creative blocks and get ideas flowing.

However, working with AI also brings a new set of challenges. While it is a powerful tool, finding the right balance between human intuition and technological assistance is not always straightforward. Designers still need to preserve their own voice and authenticity. If relied on too heavily, AI-generated content can begin to dominate the process, resulting in work that feels less personal and lacks emotional depth.

There is also the issue of skill development. Relying too much on AI can make it tempting to skip traditional learning, which is essential for building strong creative foundations. Over time, this may limit a designer’s ability to think critically, experiment, and develop original ideas.

Ultimately, AI introduces both opportunities and challenges. The key is to use it as a supportive tool while ensuring that human creativity, judgment, and individuality remain at the center of the design process.

Winning speed and money using AI

Today, AI is making design faster and more cost-effective by reducing the time spent on repetitive, manual tasks. Things like adjusting spacing and layouts, rewriting UI copy, or identifying visual patterns can now be done much more quickly. Tasks that once took hours or even days can often be completed in minutes, freeing designers to focus on more meaningful creative work.

Tools like Midjourney, and DALL-E allow designers to generate concepts, variations, and visual assets almost instantly. In the past, producing multiple design directions required significant time and effort, but now it can be done far more efficiently.

There are already clear examples of this impact. Companies like D2L Brightspace got over 100 variations of brand imagery in just a third of the usual time. Similarly, Paw-Some reportedly saved up to $20,000 while generating more than 750 unique images and many more.

Overall, AI is not necessarily replacing designers but supporting them, helping teams work faster, reduce costs, and take on more ambitious projects than before.

Downside of a generation

While AI offers many advantages, it also comes with significant downsides that are becoming hard to ignore, and that also not everybody knows of yet. One of the biggest concerns is its environmental impact. Training and running large AI models requires enormous amounts of energy, often powered by data centers that consume vast amounts of resources and contribute to carbon emission. Some tech companies in the US are increasingly buying up residential land and rezoning areas for industrial use to build data centers. People fall without water because all the servers that are being used need cooling to work properly, and in turn it uses the freshwater reserves.

At the same time, the growing demand for AI has driven up the need for powerful hardware, particularly GPU’s produced by companies like Nvidia. Random-access memory (RAM) prices have skyrocketed because demand is so high right now. This has caused components to have a limited availability and affect not only AI development but also gamers, researchers, and smaller companies.

On the other hand, this technology is so freely available that everybody is starting to feel creative and designer-like. Without really having a clue what they are doing, people start generating things and fill up the internet with bland and repetitive designs or even misleading content. This also ties into a concern of originality. AI models are trained on existing data, which can lead to outputs that feel derivative rather than truly creative.

Finally, AI raises ongoing debates around copyright and ethics. Since models are trained on vast amounts of existing work, often without clear consent, it remains unclear who really owns the output and whether the process fairly respects original creators. Which also leads to my next topic.

Copyright issues designing with AI

Copyright has been an important topic since the 18th century and has continued to evolve over time. At its core, it exists to protect creators. Designers and artists want to ensure that their work is not copied or used for commercial purposes without proper permission or licensing.

At the same time, modern AI systems like Generative Adversarial Network (GAN) have introduced new complexities. A GAN learns patterns and structures from large datasets using two components: a generator, which creates new images, and a discriminator, which evaluates whether those images look real. Through this process, the system learns to produce visuals that resemble real data. Importantly, it does not directly copy and paste pieces of existing images but generates new ones based on learned patterns, such as shapes, textures, and lighting. However, in some cases, especially with limited data or overfitting, the results can look very similar to existing works.

This has raised concerns across creative industries. In film, for example, actors have pushed back against the use of deepfakes that replicate their likeness without consent. Artists are also experimenting with ways to protect their work, such as applying filters that interfere with AI training. Meanwhile, institutions like the United States Copyright Office have stated that copyright law only protects works created by humans, meaning purely AI-generated content may not qualify.

Overall, AI is challenging traditional ideas of ownership and authorship, making it increasingly important to rethink how creative work is protected in a rapidly changing landscape.