AI-assisted design. Exploring the limits of human-machine collaboration

As the possibilities for innovative collaboration between humans and machines multiply, it is difficult to ignore the complications ahead. Widespread AI use raises important questions, concerns and challenges around authorship, ethics, and human creativity.

Our Immersive Design practice explored them further by collaborating with AI in creating our virtual studio: a space to socialise, be creative and share personal projects. We moved forward from the traditional spatial design approach to see how far we could go before reaching the limit.

Tools and design process

We started with setting guidelines for the design process - a space layout necessary for our virtual activities, translating company values and identity into keywords, and collecting visual references. Then turned them into broad instructions (text and image) and prompted AI models such as Midjourney, Stable Diffusion and DALL-E. We integrated the 2D AI outputs into our mood boards and repeated the process several times to solidify the concept. 

This technique turned into a smooth dialogue between a designer and a machine, where both sides fed each other with 'inputs' and continuously evoked new ideas.

At this stage, the speed increase in our envisioning process was remarkable. Still, current open-source AI models do not understand real physics in a complex 3D space, and we had to trust our design skills for modelling the final objects. 

To extend the output variety, we tried to delegate further. We used POINT-E to generate 3D models by creating clusters of data points in space. We were happy to discover that the imperfect random outcomes cultivated our creativity even more despite its limited precision to simple objects.

Benefits and drawbacks of AI-assisted design

The experiment made us believe this approach has as many benefits as drawbacks. And it is up to us to navigate them and find the right balance between human creativity and machine support.

Among the benefits, we appreciated the increased efficiency in generating ideas and the new perspectives that emerged from patterns we humans hardly recognised. Also, the cost-effectiveness of integrated automation for certain tasks and the enhanced personalisation of desired solutions were considerable.

On the other hand, AI algorithms lacked originality due to their reliance on existing data, which led to sensible ethics and bias issues. Its outputs were only as good as the data we fed, and biased data produces biased results. Also important was AI’s lack of quality control, refinement and customisation only humans can guarantee, especially in our design practice.

Could we have done it without AI?

Probably, but at a cost. Leveraging generative AI and parametric design tools has provided us with many possibilities beyond our reach. This is due to our inability to generate many ideas in little to no time. Our results would have been limited to our assumptions around the project, especially the subconscious ones. 

Collaborating with AI has stirred our directions into unexpected and valuable places despite, or thanks to, its randomness and “errors”. Within creating a virtual venue in Web3, AI tools radically sped up the process and widened the creative possibilities. With time, widespread utilisation, easy access, accuracy and tailored legislation, AI will become an essential element of our creative design process. 

AI tools themselves may not generate the best solution, but in our hands, these results can help reach far greater outcomes. We look forward to when the tools will be constantly refined, regulated and respectful of other creators' work.

Human creativity might not be at risk 

AI is not a replacement for human creativity but rather a tool to enhance and support it. While it has a high execution speed and infinite generative power, it has no conscious understanding of human emotions, experiences, and contexts as humans do. 

Creativity often requires a deep understanding of these complex human factors. As such, AI alone may not always be able to match the flexibility and empathy required to solve complex creative problems.

Due to its dependence on input data, it may not be able to produce truly novel or innovative ideas. On the other hand, human creativity is not limited by data or algorithms and has a chance to produce truly unique and groundbreaking work. The key is to find the right balance between human and machine support and use AI to complement rather than replace human creativity.

How to better prepare for what’s coming 

When working on clients’ projects, there are concerns about the originality and uniqueness of AI-generated outputs. If they are too similar to existing work, it could lead to poor creative output or, worse yet, plagiarism. Our responsibility is to ensure that human-AI collaboration does not damage clients’ assets and reputations.

To better understand and prepare for AI’s impact on our professional and personal lives, we must strive to stay informed about its latest development and applications. We also have to make experimentation with AI tools a good habit, collaborate with AI experts to understand its scope of application further and, most importantly, never forget the ethical implications. 

Faster, better and easier are relative concepts and don’t always mean good. Before taking advantage of the opportunities coming with using AI, we should always ask ourselves if we need it and at whose expense. 

The AI-assisted design possibilities are well beyond our current understanding. With careful experimentation and responsible practice, we can harness its potential for society’s benefit, as we like to think we have - at times - done in the past.