Day Fourteen: Automation of Secondhand
Take a listen to the following audio. How can current types of automation such as AI, and especially generative AI, impact secondhand economies?
More on Automation
There is a long history of fashion and automation, with mechanical and automated work historically displacing the work of artisans and craftspeople in the UK ad Industrialized countries. Together with colonial trade, the transatrlantic slave trade and the institutionalization of plantations, plu displacement of non-automated poorly paid garment work to the global south, this automation has accelerated overconsumption and capitalist accumulation. In other words, this automation, which in itself is not necessarily damaging to the planet (although some academics might question this), has contributed ultimately to many negative aspects of the fashion industry in combinatin with sociopolitical structures in place.We’re going to ask now: How can current types of automation such as AI, and especially generative AI, impact secondhand economies?
We previously talked about the ways algorithmic systems in platforms create information asymmetries, invisible work, and foster a kind of collective sensemaking from communities of resellers trying to understand how to use the systems in their benefit. Communities of secondhand workers like resellers are diverse, displaying a range of opinions and approaches to their businesses and relationship to secondhand economies. This means their responses to automation are also diverse: finding resourceful responses to the constraints of platforms. An adjacent community of third party providers (which include human laborers in the global south, and click pressing bots) have also become important actors of secondhand economies, seeking to cash on the value latent in secondhand.
With the popularization and corporate hype around generative AI, platforms are incorporating tools that automate some of the tasks done by resellers: photo staging (in addition to background removal which was already used without gen ai, Ebay also has incorporated this functionality), auto description writing based on machine vision and eBay datasets. Third parties and bot companies are also looking to tap into the AI hype, building tools with similar purposes.
AI is also built on invisible labor, requires powerful systems and stolen data, pretends it is more automated than it acyually is, and has terrible environmental effects.
Timnit Gebru and Meredith Whittaker have consistently underscored the ethical complexities and societal implications of artificial intelligence (AI) and machine learning (ML). Timnit Gebru has pointed out that "AI systems are not neutral; they are imbued with the biases of their creators and the datasets they are trained on." This often leads to technologies that "replicate and exacerbate the very biases and power dynamics present in our society," disproportionately affecting marginalized communities.
Meredith Whittaker has echoed these concerns, stating that "AI is a mirror that reflects the values of the society that creates it." She emphasizes that the development of AI is frequently driven by powerful corporate interests, leading to a lack of accountability and transparency. Whittaker warns against the "uncritical adoption of AI technologies," highlighting potential harms like surveillance, privacy breaches, and labor displacement.
Both scholars also draw attention to the often overlooked but crucial role of data workers and labelers in the AI ecosystem. These workers, who painstakingly annotate and prepare the data used to train AI models, are essential yet largely invisible in the narrative of technological progress. As Gebru notes, "The invisible labor of data workers and labelers is the backbone of AI systems, yet their contributions are often undervalued and poorly compensated." This workforce, frequently based in the Global South, performs the tedious and challenging work of ensuring that data is accurate and representative.
Whittaker adds that this lack of recognition and fair compensation not only raises ethical concerns but also affects the quality of AI systems, as these workers are crucial in defining the datasets that shape AI's outputs. Both advocate for recognizing the vital contributions of these individuals and ensuring fair working conditions, as they are integral to the creation of these technologies.
Examples