Last week, when I attended this year’s VOICES – BoF’s major annual gathering for thinkers and industry leaders – I had the opportunity to talk to people in various roles in fashion. Of course, being a technology reporter, a lot of those conversations were about innovation.
As I talked to people at brands, intelligence agencies, and fashion education, one technology came up again and again, and it wasn’t virtual reality or the metaverse or NFTs. It was AI.
Artificial intelligence is a concept that dates back several decades, but in recent years advances in machine learning and computer vision have allowed it to have a real impact on businesses. Today, the most tech-savvy fashion brands use it for practical tasks such as forecasting demand and setting prices. But as progress continues to accelerate, the range of possible applications continues to expand. People in fashion are thinking about where it could go and what possibilities it could offer soon.
With VOICES, part of this interest may be due to timing. Right now, there’s a lot of excitement in the tech world around text-based AI image generators like DALL-E, Midjourney, and Stable Diffusion. A use case that seemed to draw a lot of interest from VOICES participants, which I’ve also written about here, is how generative AI can be used in design. A few people said they thought it seemed inevitable that AI would eventually play a role in the creative side of fashion. (No one went so far as to guess that it would completely replace humans.)
And that wasn’t the only new application of AI that emerged. While guests were still interested in topics like the metaverse and NFTs, it seemed that AI more than any other technology would have the biggest impact in the near future.
For example, another person I spoke to suggested that computer vision could soon enable robots to sew clothes—a feat now attempted by a few companies without much success because robots have difficulty controlling and manipulating flexible, stretchy fabric. When I pointed this out, the person commented that computer vision is improving so fast that technology today enables feats that were not possible a year ago, and even more will be possible in two years’ time. (Of course, people have been saying that for years, too.)
The rapid developments in the field of AI were also the subject of one of the presentations at VOICES this year. Mo Gawdat, an author and former chief business officer at Google X, which describes itself as a “moonshot factory” that uses technology to solve some of the world’s toughest problems, spoke of several “inevitable things,” including that against the end of this decade the smartest creature on earth will not be human. It will be an AI.
There is still debate about how “intelligent” artificial intelligence really is. Is it capable of real learning and not just doing limited tasks assigned by humans? According to Gawdat, it is. He pointed to the “cat newspaper,” where Google explained in 2012 how a network of computers it trained on YouTube videos finally identified what a cat is without any human input. Similarly, in 2016, Gawdat said Google had a collection of robot grippers that it directed to pick up various objects without instructions on how to do it. On a Friday afternoon, a grab managed to pick up a yellow ball. On Monday all grabs were picking up yellow balls. Two weeks later, they were able to pick up every object available.
Whether or not you believe Gawdat’s view of the future, which envisions a time not so far away when the world is completely reshaped by machine intelligence, you can believe that AI will continue to get better and we can use it for a greater number of tasks. A robot using AI to learn how to handle and sew fabric doesn’t seem so far-fetched.
There’s a lot of buzz online right now around ChatGPT, an AI tool created by DALL-E’s developer, OpenAI, that can compose elaborate text responses to user prompts. “ChatGPT is simply the best artificial intelligence chatbot ever released to the general public,” one New York Times tech columnist recently stated. It has its flaws. It doesn’t seem to understand the meaning behind its responses and may contain information that is flat out wrong. But it’s also not hard to see how such a tool could be used, for example, to write marketing texts (for now, at least, to be reviewed by a human being).
As AI becomes more common, it probably won’t eliminate every job it touches, but it will change them, changing the jobs of the people involved and the skills required. But technology is always changing the way people work.
The feeling I felt most with the VOICES guys I spoke to about AI was not fear or panic. It was something like curiosity. No one knows exactly how it will develop or how it will change the day-to-day of fashion in the next two, five or ten years. But that it will continue to advance seems inevitable, to borrow a word from Gawdat, and fashion will follow.