Introduction

“Innovation” has been at the forefront of modern technological products. For many companies, innovation becomes a motivational quality, a marketing scheme, and a means to justify any negative impacts their products may produce. This has been especially apparent for modern artificial intelligence technologies, specifically machine learning methods such as ChatGPT by OpenAI. This project seeks to examine the ever-so-leaning balance between seeking innovation and addressing the plethora of negative effects manifested through such process. Specifically, we attempt to provide a meaningful solution to the following questions:

To what extent has the development, deployment, and maintenance of computational methods in the recent decades, particularly machine learning algorithms affected the global environment? How should technologists ethically balance the scale between pursuing innovation and the subsequent environmental impact these processes are associated with?

Through the process, we will also explore previous work that addresses the topic by researcher and academic, Kate Crawford, and make efforts to provide an accurate and provoking visualization of the magnitude of environmental impact examples of the aforementioned technologies, such as ChatGPT, have with a goal of creating and spreading awareness to the tech community and, as an extension, the general public.

Atlas of AI

“Atlas of AI” by Kate Crawford provides a critical and comprehensive examination of artificial intelligence’s social, political, and economic impacts. She argues in her work that AI technologies must not simply be seen as innovative tools but also as an incredibly intricate and complex system that involves the utilization of natural and human resources. So much so that without such involvement, these systems would not be able to exist or function in any meaningful way.

We will primarily focus on the environmental aspects of Crawford’s critique. In the chapter named “Earth”, Crawford highlights the undeniable connection between the origins of Silicon Valley and by extension the tech scene in Northern California and the extraction of minerals such as lithium in nearby mines, many of which have been depleted of its resources. She also emphasizes the importance of these metals and minerals as an integral part of the development and deployment of AI systems as, for example, “rechargeable lithium-ion batteries are essential for mobile devices and laptops, in-home digital assistants, and data center backup power”. More importantly, Crawford emphasizes the idea that the ignorance of the supply chain is woven into the fabrics of capitalism and the there exists a strong sense of deniability of unsustainable and harmful extraction practices on the part of many large tech companies. Worse still, these issues extend beyond being purely domestic. Similar patterns of extraction have been identified in China, Indonesia, and the Republic of the Congo as companies move their supply chains overseas. Crawford’s dialogue on the environmental detriments AI systems produce portrays a incredibly thought-evoking narrative behind the highly innovative and attractive facade tech companies characterize their products with.

Visualization

The following visualization tool aims to provide a rough estimate of ChatGPT’s energy and water expenditures in real time starting from when the user loads the page.

Takeaway

It is important to recognize that the goal of this project is not to undermine the efforts and time many tech companies such as OpenAI have put forth into developing and producing these innovative products and technologies. This project does, however, aim to spread awareness and engage dialogue within the tech community regarding the environmental impacts these products may carry. If you’ve happened to stumble across this page, I highly urge you to apply the knowledge you’ve acquired in the aforementioned manner so that we may hope to balance the scale between pursuing innovation and saving our home.