AI forecasting and the future of work: we need more responsible and inclusive narratives
By: Tessa Boumans
In their recent Business of Fashion podcast, technology correspondent Marc Bain joins Imran Amed to discuss how generative AI (such as Chat GPT and DALL-E) might impact the fashion industry. Although they touch upon both the risks and benefits of artificial intelligence, they leave out a crucial critical systems perspective, and in doing so ignore the hottest potato in the discussion: How will AI impact all those women in the Global South that make our clothes? In this short blog post, I’ll try to unravel some of the dominant power dynamics that lie behind the narratives in the current AI and fashion discussion and reflect on what we need to aim our analyses at in the future.
Let me first mention the important points that they did address in the episode. To summarize Bain, there are several key elements to the AI and Fashion discussion. First of all, generative AI has the potential to erase or at least drastically alter the jobs of fashion designers, as well as more ‘behind the scenes’ jobs such as content creators or customer service employees. Second of all, we need a better understanding of the extent to which generative AI is copying intellectual property and can actually generate new, desired styles. Although, as Bain also mentions, copying is already happening at large in the fashion industry (AI would just make it easier), and creativity is often “romanticized”. Often, it’s not so much about creativity as it is about efficiency and having a market to sell to. Ultimately, these futures are not set in stone. As Bain mentions, it comes down to whether brands will be able to capitalize on the available technologies. Most of what it will come down to is if they can make money of it.
This brings us to a third important issue that was raised in the episode: the reproduction of bias. As Bain rightfully mentions, AI doesn’t know what’s right or what’s wrong. Rather, it just reproduces what it has seen in the past. As he says: “the way AI works is it’s trained on past information. So whatever biases, whatever problems exist in that past information: it can repeat them. So, that could mean anything. It could mean using racist language because its finding that stuff on the internet. It could mean coming up with an ideal dress or something that is an ideal design for a slim white woman and not every range of body types. So, all that sort of stuff, anything that has been an issue in the fashion industry in the past, that is something to look out for.” From this quote, we can identify that Bain — albeit indirectly — mentions systemic issues in the fashion industry such as racism and sexism. The trained mind (a.k.a. anyone familiar with prevailing issues in the garment industry) will also recognize issues of colonialism and capitalism. Not coincidentally, these systemic inequities are fundamental barriers to achieving decent work for all in the garment industry. And the ways in which AI and other advanced technologies are likely to play a role in maintaining this status quo are plentiful.
Bain’s comments might seem like a critical note, and in a way they are, but these critiques are not comprehensive enough. If we want to prevent AI from reproducing the current system of mass exploitation and injustices against garment workers in the Global South, we need to actively intercede. In order to establish a sustainable transition for the industry, we first need to understand system dynamics of AI and work, and then identify leverage points to intervene. At the same time, we need to be concrete and transparent about our ideas of a sustainable future: because if we intervene, what do we want to achieve exactly? And who has been given the voice to express those desires for the future? I’ll tell you one thing: it’s generally not garment workers.
The current discussion around AI and fashion is — again — too focused on perspectives and the context of Western brands, consumers, and workers. In that sense, this particular podcast episode is representative of the larger discourse that prevails around tech in Global Value Chains. Naturally, individual actors in these industries are also a victim of internalizing dominant capitalist, colonialist, sexist, and racist narratives. But as discussers with a stage, we also have a responsibility to break from those narratives in support of the globally marginalized. In addition, it is worth mentioning that of course the availability of jobs for young American designers is important. Of course, the issue of intellectual property and creativity is important. But in an industry that is characterized by structural human rights violations in specific spatial parts of its value chain, we also need to be critical about what topics we prioritize in our discussions around AI and fashion. I’m not advocating that we should stop discussing the impacts mentioned in this episode, but it would benefit the sector if these discussions were never again held without at least a mention of the labor rights of garment workers in the Global South. Considering the scope of violations in the sector, that seems like a small question.
