DESIGN DISKURS
Generative AI opens up creative possibilities – but it raises questions about human-crafted design. While AI can offer efficiency and streamlining, the essence of design lies in our ability to surprise, innovate, and bring a human touch to our creations.
In the last few years, Artificial Intelligence (AI) has become a common term and is increasingly shaping our everyday lives. With ChatGPT, we are currently witnessing an impactful evolution – the one of generative AI. Generative AI is a subset of AI technology that utilizes machine learning and deep learning to create new content based on what it has learned from existing data. Generative AI models go beyond replication by generating diverse and contextually relevant outputs in various formats such as text, code, images, audio, and video.
The process of Generative AI involves training a model on existing content, known as training data, and then using the model to predict and generate new content based on a given prompt. For example, advanced language models like PaLM or Pathways Language Model and LaMDA can ingest vast amounts of data from the internet and provide comprehensive responses. The amount of data used to train a model in Generative AI can vary depending on the specific model, the complexity of the task, and the desired level of performance. There are often billions or trillions of data points, to train robust foundation models that can effectively generate new content. With 175 billion parameters that capture a wide range of language patterns, ChatGPT is the most prominent large language model at the moment (Alarcon, 2020).
“To keep AI models on the rails, human oversight remains essential.”
Training enables a model to learn the underlying patterns and structures within data. By leveraging deep learning techniques, which involve artificial neural networks with interconnected layers of neurons, Generative AI can process complex patterns and generate new samples that resemble the training data. Foundational models can be fine-tuned for specific tasks and domains, often requiring significantly less data compared to the initial training phase. To keep AI models on the rails, human oversight remains essential. This ensures responsible behavior and unethical content can be removed, leading to continuous improvement through reinforcement learning from human feedback (Heaven, 2023).
By pushing the boundaries of what AI can achieve, Generative AI opens up creative possibilities, causing concern in industries like design, art, and advertising. Diffusion models in particular like Dall-E 2, Google’s Imagen, Stable Diffusion, and Midjourney are a headache for us creatives. These models have the power to generate any image that you can imagine from text prompts, including the photorealistic, the artistic, and the futuristic in seconds instead of hours.
Three Approaches to Design and AI
Examining the dynamic between AI and design, raises questions about the value of human-crafted design in the age of Generative AI. In March of this year, the Future of Life Institute issued an open letter advocating for an immediate six-month pause on the training of AI experiments. In contrast, a report from Goldman Sachs' 2023, which predicts that the integration of AI tools into businesses and society could lead to a 7 percent increase in global GDP (nearly 7 trillion Dollars) and raise productivity growth by 1.5 percentage points over a span of 10 years. In addition, after only two months of release, ChatGPT had already been used by over 100 million users (Milmo, 2023). As designers, we must acknowledge that we cannot ignore, nor stop, or pause the trend of artificially generated content, as it progressively integrates into our daily work.
The relationship between design and AI can be categorized into three distinct areas: Design for AI, Design with AI, and Design by AI. Understanding the unique opportunities and risks of each approach is crucial. Designing for AI allows us to shape user experiences and ensure intuitive and beneficial AI applications; however, the risk lies in users potentially losing trust in AI systems if the designs are not executed effectively. Design with AI presents opportunities for collaboration, leveraging the strengths of both humans and machines, but there's a risk of over-dependence on AI suggestions, potentially stifling human creativity and originality in the design process. Design by AI can offer automation and efficiency, freeing us to focus on creativity. In the worst case, automation could lead to designers being disengaged from the creative process, resulting in less understanding and control over the final design outcome. By grasping these distinctions, we can navigate the landscape of AI and harness its potential to enhance our creative work without compromising the value of human ingenuity and surprise.
“AI is limited to reusing what it has learned in an uninspired manner.”
Considering these approaches, it is evident that our job as designers has become both easier and more challenging. It will become harder to stand out from the vast expanse of generated content and striving for distinctiveness. But thanks to AI, we can focus more on the complex aspects of our work, allowing us to explore problems in detail and find intelligent solutions. Let's be honest – surprising and creating new trends is the hard part of the creative business. While AI can streamline certain tasks, its capabilities are limited to reusing what it has learned in an uninspired manner. Relying heavily on Generative AI makes design repetitive and runs the risk of making us replaceable. Design, in its essence, goes beyond replicating existing patterns; it should strive to surprise, to introduce the unexpected, and to shape unique trends. This is essential for exceptional and original design. It boils down to mindset – an unwavering determination to surpass everything out there. This, indeed, is demanding. And yes, in the end AI will learn from us again. Absorbing our knowledge and experience. Reuse what we are creating.
“AI serves as a tool to augment our creativity rather than replace our individuality.”
Our creative energy, originality, and ability to think beyond the familiar is what truly set us apart. Critical thinking and evaluation are essential in this data-driven landscape because design is not truly complete until it is put to use by someone (Laurel, 2001, p.102). This should remind us that design must always address the needs, preferences and experiences of those who interact with it. While AI can augment design processes, it is the human touch that brings the element of surprise, originality, and user-centricity to design. We should take a critical look at Generative AI, apply critical thinking, and ensure that AI serves as a tool to augment our creativity rather than replace our individuality.
Design as Bridge between Immaterial and Material Values
This year, our DDC WAS IST GUT competition explores the values of design. The competition raises fundamental questions about the role of design and its capacity to tackle real-world challenges. It encourages designers to challenge conventions and create meaningful content, products, and processes that generate value and contribute to a better world. Material values are those examined by economists, while immaterial values are described by philosophers and social scientists. Felix Kosok emphasized in his DDC DESIGN DISKURS the role of designers as experts in bridging the gap between these two types. He described that economists and philosophers alone cannot answer the question of the connection between different values. Designers are the ones who have the knowledge and skills to translate abstract ideals into tangible structures. They play a crucial role in shaping value by creating meaningful communication, developing products that add value, and designing spaces that are in tune with societal transformations. By addressing social, ecological, political, practical, economic, and aesthetic aspects, design serves as a value-creating process that raises essential questions and seeks new answers for the betterment of our societies.
By highlighting the importance of immaterial values alongside the material ones, WAS IST GUT points out design as a bridge between the two realms and inspires reflection on the values that shape our design practice. With the adoption of AI doubling since 2017, important questions arise: What are the real differentiators of human-crafted design? (Chui et al., 2022). How can we create value/s through AI? And how can we use Generative AI in our work for the better? We must recognize that, despite working alongside machines, we remain the creators and authors of what emerges.
“Human designers bring creative intuition, empathy for users, iterative problem-solving abilities, attention to detail and craftsmanship, and a commitment to ethical and social responsibility.”
Generative AI models have created new ways for designers and professionals to create content. In 2022, Thomas Davenport and Nitin Mittal described four significant advances this technology brings to the design process: automated content generation, increased content variety, personalized content based on user preferences, and improved content quality. So the integration of Generative AI in design processes can provide significant benefits. While it offers opportunities for efficiency and diverse output, it is the distinct combination of human skills and core values that sets human-crafted design apart. Human designers bring creative intuition, empathy for users, iterative problem-solving abilities, attention to detail and craftsmanship, and a commitment to ethical and social responsibility. These skills enable a human-centered approach where designs evoke emotion, seamlessly blend art and functionality, and produce mature and lasting results that also consider their impact on our environment. As we apply Generative AI to our design work, the challenge is to ensure that the intangible values of design are woven into the tangible creations of generative AI. In this regard, human-created design plays a central role in bridging the gap between material values (created by AI) and the immaterial values that are the essence of design.
Ethical and Legal Considerations
We are still in the early stages of AI and Generative AI. In the coming months, we can expect a flood of new tools and models from diverse sources – from large companies to emerging startups and from tinkerers in small garages – who are developing groundbreaking ideas. These innovations are no longer just coming out of Silicon Valley, but are emerging around the world. Some ideas may seem unusual at times. One example is the dating app from Teaser AI, which creates an AI version of its users that mimics their personality and conversational style based on their profile.
A major dilemma for creators is the tension between copyright and Generative AI. When training AI models for image generation, the data is broken down into its smallest units and then reassembled to generate new images. This leaves the originators vulnerable and with no legal protection. (Lobe, 2023). “Have I been trained” is an exciting counter-trend in this legal and ethical discussion. The tool allows users to explore a vast repository of 5.8 billion images that have been used to train renowned AI models and helps to check whether their own images were unknowingly part of the training process. It then provides an opt-out function, similar to classic newsletter unsubscribe links, that allows users to decide whether or not their work can continue to be used for AI development. Such an approach raises important questions: Should we regulate AI? What laws do we need regarding AI? Are there ways to make consent mandatory before data can be used for training?
“For designers especially, the copyright discussion is intriguing.”
According to the Boston Consulting Group (BCG) “AI at Work” report, 79 percent of respondents from 18 countries, including executives, managers, and employees, expressed the need for specific AI regulations. For designers especially, the copyright discussion is intriguing. Currently, intellectual property laws revolve primarily around copyright protection for creative works. This can answer the question of who should be considered the creator or copyright owner when Generative AI is involved – the AI system itself or the human user.
The European Parliament is currently defining the “EU AI Act” which should become the foundation for a fair environment for AI development. The plan is to classify AI applications based on their risk to users and then regulate them accordingly. Initially, Generative AI did not play a role in this. It was applications such as ChatGPT that brought it to the forefront, which is why additional regulations are now being created. They want to require labeling of AI-generated content and the disclosure of copyright-protected training data. The upcoming EU AI Act is an opportunity for designers to advocate for their position and work towards fair treatment of their works. This can form the basis for a comprehensive and responsible approach to AI-generated content that promotes ethical and sustainable AI development.
As designers, we find ourselves at the forefront of a rapidly evolving landscape where AI and Generative AI are becoming increasingly integral to our work. Three distinct approaches – Design for, with, and by AI – present us with new opportunities and challenges. While AI can offer efficiency and streamlining, the essence of design lies in our ability to surprise, innovate, and bring a human touch to our creations.
“We must find a balanced approach where AI serves as a supportive tool rather than a substitute for our own ingenuity.”
The significance of human creativity in preserving originality and unexpected elements in design is essential. As we navigate this technological transformation, we must find a balanced approach where AI serves as a supportive tool rather than a substitute for our own ingenuity. We should embrace AI's potential to augment our creativity without losing sight of our design responsibilities and the needs of our users.
Understanding the distinct opportunities and risks presented by Generative AI is vital. As we explore the possibilities, let us remember our purpose as designers – experts in bridging the gap between immaterial and material values. By fostering a mindset of unwavering determination to surpass the ordinary, we can harness the potential of AI while upholding the irreplaceable value of human-crafted design.
