AI's participation in the operation of the fashion industry is not a day or two.
From supply chain automation systems, to big data and cloud computing that can assist business decisions, to communication and interaction in the digital marketing process... In the past, intelligent technology has transformed the fashion industry in many dimensions. The evolution of artificial intelligence to large-scale language models such as ChatGPT, and its impact on more interactive environments in the future can be said to be revolutionary.
But while AI continues to expand the boundaries of human behavior and efficiency, infinitely close to "people", where is the creative boundary between AI and designers? Creativity, itself is based on the perception and feedback between social relationships, relying on humans to give data, set rules, passively generate results, how far can AI go in the creative world?
In the summer of 2021, Yu Yimeng, a young teacher who currently teaches at the School of Design of the Central Academy of Fine Arts and is the convener of fashion design, led students to participate in the DeeCamp artificial intelligence training camp, one of which is "clothing design generated based on AI dialogue", which began to pay attention to and explore the "AI+human" working mode in fashion design. After introducing AI technology as a design tool in the course, Yu Yimeng said that AI, as an auxiliary tool, can quickly generate text and image references, greatly improve the efficiency of design and creation, and stimulate more creative possibilities.
Fast forward to 2022, the algorithm of artificial intelligence continues to mature, and the iteration speed is also very fast, and the subdivision of application fields is becoming more and more "specialized". Yu Yimeng mentioned that Dalle.2, launched by OpenAI at that time, caused great repercussions in the field of art and design. Recalling her experimentation, she says: "It was a pleasant surprise because the results were similar to my style, but there were endless iterations, and the subsequent Mid Journey was more in-depth in terms of precision and detail. It's just that the results are sometimes more 'Mid Journey' in the style. ”
Wang Sicheng, who works in visual arts, also pointed out that AI improves efficiency, and ChatGPT's ability to write code helps our content producers reduce work pressure, such as writing code for Unity can be used for modeling, etc., which greatly reduces the pressure of repetitive work. It can also help him arrange and combine some visual language as a reference for creation.
"It's also interesting to open your mind," Wang told WWD Chinese.
According to Zhao Shiji, a mentor in the JALAB lab of Tongji University and co-founder of the UNI jewelry designer platform, AI is more accurate in assisting users to analyze matching and preferences, and even assisting brands in market analysis, which is what she is very optimistic about.
In addition, for Zhao Shiji, as well as the founder of the designer brand Tuyue, contemporary artist Yang Yingying, the founder of the designer brand PH5 and many other design researchers or creative workers, AI is more accurate in some ergonomic related calculations, and it is currently more suitable for the design of some standardized clothing.
However, just as intelligent technology has brought changes and recognition to the fashion industry in multiple dimensions in the past, a series of concerns and discussions about ChatGPT, is it possible for AI to replace designers?
In fact, through many studies, as well as the practice of designers and artists, it has been shown that the boundary between AI and human designers is still clearly visible.
As a study by Busan National University in South Korea points out, the biggest difference between the two is the uniqueness and originality of human design, which often comes from personal experience. Assistant Professor Yoon Kyung Lee proposed that AI technology can improve design efficiency, allowing designers to focus on difficult creative work.
For independent designers, artists, and young teachers engaged in the source of the creative industry, although the application provides extremely powerful convenience, based on the current degree of application and working mode of AI in the fashion and art industry, the creativity of designers is still difficult to replace by AI.
Wang Sicheng expressed to us the excitement of starting with image content generation AI programs such as GAN and Stable Diffuion, and then slowly learning about ChatGPT and AIGC, "like getting a new tool and a new toy," he told WWD Chinese. Regarding the current artificial intelligence technology, he said: "We have been expecting AI to be more 'human'. But at the moment, I still can't make the content produced by AI become a work of art on its own. ”
Through his understanding, so far, AI-generated content and ChatGPT, or the collation and processing of past content and information, it is difficult to produce new artistic visual innovation, and its homogeneity will soon be discovered after experiencing the initial excitement. For him, AI technology is also in the role of a tool, and he probably knows what the results will be.
In response to the assumption that AI "inevitably" cannot replace humans in the field of design, Yu Yimeng believes that this is the case from the current point of view. "At the beginning, I will hear some discussions of artificial intelligence on designers' 'dimensionality reduction strikes', and I think that although artificial intelligence generative design will have surprises, there is still a distance from replacing or even surpassing human designers.
She also pointed out that behind AI is algorithm, is code, it is a tool, its advantage is to learn a large number of data sets in a short time, and then according to the rules and keywords set by humans to generate results, human designers do is to give data, set rules, do screening, do iteration, make decisions, the final product is the product of human brain thinking.
At present, in her pre-creation stage, she will use artificial intelligence technology to find inspiration, reference, or atmosphere construction, but will not directly use the results generated by artificial intelligence technology, but need to post-process.
Therefore, it is certain that artificial intelligence technology can indeed "imitate" the designer's creative model to a large extent, and make a lot of iterations on past styles and existing materials, but it is also limited to "imitation", let alone the transcendence of creativity. An important core of creative workers is transcendence.
As Wei says, "To a large extent, very much we want to bring freshness to consumers, and we need to design clothes that they haven't thought about, haven't seen." Therefore, she pointed out, as far as the current understanding of AI is concerned, it cannot replace human creativity.
At present, Zhao Shiji also mainly uses artificial intelligence technology to initially assist in the creation of models and sketches, but will not rely too much. She also told WWD Chinese edition: "If you want to talk about art and original design, people are still the original core of the work. ”
"I'm not sensitive to technology per se, so I've always thought of AI as a tool for outward exploration. I'm not pessimistic or optimistic, it's a state of pure wait-and-see," Yang Yingying told WWD Chinese.
In her work with the design team, she learned that AI can assist creativity, but there are currently more "wrong" samples, but through continuous training and "feeding", errors can be reduced as much as possible. She told us: "Because we don't do standardized things, in a way, manual work is still more efficient." ”
Although for designers and artists at the source of creativity, there are different degrees of barriers to current artificial intelligence technology.
However, just like the speed of development of technology, the degree of iteration and deepening of AI itself. In the future, how far AI will go in creative work is still full of unknowns and possibilities.
On the one hand, as we have already seen, at the technical level, in terms of channels, information data collection and computing, etc., artificial intelligence technology has brought subversive and positive changes to the operation mode of the fashion industry in the material world, and the scope of application is becoming more and more extensive.
On the other hand, artificial intelligence technologies that once relied on humans to set rules, input information data, and passively generate results, when they adapt to more rules and larger information data, how powerful this "passive" force is, determines its participation in human creative activities and the depth of social civilization symbiosis.
In Wang Sicheng's view, if you look at the long term, the ultimate goal of AI is to become a "person". In the beginning, it was an elementary school student, a junior high school student, and now ChatGPT has shown the ability of a college student. If it continues to evolve in some segments, perhaps AI can also be trained into an innovative model.
He mentioned his work called "Ten Thousand Resistance", whose vision mainly uses GAN technology to supplement the breakpoints of statue sculptures in the process of cultural transmission, similar to the changes in the evolution of animals. "I feel like it inspires me on a conceptual level and gives me room for fantasy. Although the content created by AI itself cannot be a work of art on its own, it can provide methodological and perspective inspiration in contemporary art creation. ”
"I'm not proactive, but I'm definitely open", according to Yang, design and creative work often requires brainstorming, and AI is undoubtedly a good object.
When it comes to the question of whether artificial intelligence technology is destined to be a layman for designers with avant-garde creativity and unique styles, Zhao Shiji also said that it is not always true. "AI technology can also help independent designers broaden their avant-garde style in a horizontal dimension and produce more interesting design blueprints."
Although at present, the advantages of artificial intelligence technology are still to help designers improve design efficiency, provide design inspiration, reduce trial and error and waste of resources, and the final result is that human decision-making is made after thinking.
But on the other side of the coin, in her opinion, "the unexpected" is also another advantage of AI technology. Colleague she also believes: "In fact, proficiency in applying AI and training to achieve satisfactory results is a very time-consuming process that requires a lot of experience." In addition to working patterns based on image generation, building your own text library and text logic is crucial and worth exploring. ”WWD
Written by Du Mingwei
Edited by Li Jun
Image source: Designer, artist official courtesy