Title | Summary on “Anatomy of an AI System, by Kate Crawford, Vladan Joler (DMS220) |
---|---|
Author | angel rivera |
Course | Machines, Codes and Cultures |
Institution | University at Buffalo |
Pages | 1 |
File Size | 47.9 KB |
File Type | |
Total Downloads | 20 |
Total Views | 130 |
Summary on “Anatomy of an AI System, by Kate Crawford, Vladan Joler...
Summary on “Anatomy of an AI System, by Kate Crawford, Vladan Joler
In the essay written by activist scholars Kate Crawford and Vladan Joler, they critique AI systems, Amazon’s Echo in particular, illustrating the origins of the AI systems and technology we know today, and pointing out various processes and events that are not known to the average consumer. For instance, in chapter 4 of the document, it explains how there is a connection between the literal excavation of materials from earth and AI, or the “cloud”. Chapter 5 examines differences between certain types of consumer technology. For example, Amazon's echo may rely on verbal commands, questions and responses in order to train and get smarter, but human knowledge and capacity may also be used as well. Chapter 6 focuses on how the Amazon Echo perceives human interaction. In many ways, humans act as consumers, a resource, and a provision of labor for the Amazon Echo. This in turn helps optimize the Echo and increase its learning capabilities. After reading chapter 11, we learn about supply and production processes of the physical inner workings of AI systems behind closed doors. Usually not transparent, there is evidence that reflects the often poor and dangerous working conditions and wages of laborers that usually excavate resources that eventually go inside the technologies such as Echo. The supply/production chain is very complex. Actually it’s so complex that it took semiconductor chip manufacturer Intel over four years to understand its own supply chain, often involving chains within chains spreading out to over 100 countries. Then, in chapter 13, the authors discuss the relations of human workers between robots. They even show an Amazon-owned patent for a system that will ultimately have a human laborer inside a metal cage with different cybernetic add-ons that can also be moved throughout the warehouse via a motorized stock-shelving system. Here, the worker becomes one with the “machinic ballet”. Finally, chapter 17 shows us how important data on AI systems, such as energy consumption is usually non-transparent- but maybe it is because the average consumer either does not want nor have time to learn about the origins of the cloud? Anyway, Crawford and Joler go on to explain how consumers have become accustomed to smaller devices such as phones and the Echo in their homes, they are not effectively exposed to the idea that there are larger, smarter systems that play a role in artificial intelligence....