The very first theme of our proposed class takes a step back from data mining specifically and looks at technology as a whole. In order to better understand data mining and the issues that arise from it, we have to understand more about technology and its place in society. What exactly is technology? It’s an ambiguous, “slippery” term (MacKenzie & Wajcman, 1985), since technology takes many different forms and uses; it’s hard to pinpoint a single definition of technology. Superficially, it would appear that technology is just technical hardware or software, the object itself. This perspective is narrow minded and leads to a theorizing of technology as deterministic. In this frame of mind, humans have no agency over technology, and its development sets us on a predetermined path into the future. Taking a deeper look at technology, however, debunks this theory and shows that we are not at all just helplessly along for the ride.
The first step of our proposed curriculum is to reframe students’ view of technology as a socio-technical system. First, we return to the question of what is technology? It is not just the technical object itself, but rather a combination of objects, activities, and knowledge (Matthewson, 2011). Take any piece of technology no matter how simple it may seem, a TV remote for example. To successfully use this remote requires objects: the remote itself as well as a TV, activities: pressing of the buttons and pointing it towards the TV, and knowledge: of how the buttons work. The same can be applied to data mining: the objects required include algorithms, computers, internet, etc. The activities required are people creating algorithms, users uploading data online, and the algorithms then collecting and sorting that data so it can be interpreted. Finally, humans have to be knowledgeable enough to train the algorithms, interpret the information output from them, and know how to use that to their advantage. From these examples it is clear that a technology as an inanimate object alone does not have any impact on society.
The missing piece to the puzzle as we have seen are humans and society at large. Political, cultural, and historical context all matter when discussing technology. It influences how a certain technology is used, or if that technology is even adapted at all. Why did the British navy make a huge push to convert their ships to run on oil rather than the traditional coal in the early 20th century? Because cultural tensions with Germany and a rising fear of war lead them to look for ways to make their ships more efficient. (Yergin, 1991). The VHS became more prolific than the Sony Betamax despite the latter being technologically superior because the American cultural and economic context paved the way for it to be adopted (Mathewson, 2011). This shows us that “technologies cannot be abstracted from the environments which they help to create.” (Matthewson, 2011). This completely debunks the theory of technological determinism where technology alone determines the course of humanity.
Technology and society must be viewed together; one does not have a unilateral impact on the other, rather they both influence each other. Humans create the technology, and it is shaped by the environment in which it is brought up; in turn, the technology affects how we live our daily lives and by extension transforms our society. Today we live in the age of big data where information about us is being collected from almost all facets of our lives. We have to recognize this as the cultural, economic, and political context in which data mining is being used. Cultural in the sense that we are doing more and more activities online (this trend is accelerated by the COVID outbreak), economic in the sense that companies can capitalize on our data to make their businesses more efficient and profitable, and political in the sense of the debate at the Federal and state level on how to regulate data collection and mining. The rise of big data is the environment which led to the use of data mining, and as we will see when we discuss some of the surrounding ethical issues, data mining is also having a big impact on us as well.
The main focus of our proposed curriculum is to analyze some of the ethical issues that result from data mining, but it is important to first lay the groundwork of data mining as a socio-technical system. Students may enter the class with the misconception of technology solely as an object and they may wonder how ethical issues could arise from an inanimate artifact? The logical conclusion is that it is the humans themselves, by controlling and directing the data mining process, that are the cause of these ethical issues. This perspective is not technological determinism, but in fact the exact opposite: social determinism (Matthewson, 2011). This theory focuses solely on us, the humans. We have complete control and our society unilaterally shapes technology. In essence, technology is the effect of our societal progression, not the cause (Matthewson, 2011). Just like technological determinism, however; this theory is misguided. An extreme social determinist refuses to recognize the impact technology has on society. How would data mining have come about if computers and the internet had not made data so prolific in our world that a way to organize and structure it was necessary? The correct way to analyze data mining is somewhere in the middle.
By viewing it as a socio-technical system, students will recognize that data mining and society cannot be analyzed separately, and that both influence one another. They will take into consideration the cultural, economic, and political environment in which data mining is being used. They will know that data mining is neither good nor bad nor neutral, since humans create technology and instill within it their own values. They will view it not as an inanimate artifact, but rather as a combination of objects, activities, and knowledge, where humans play a prominent role. If we want students to fully understand the ethical issues surrounding data mining, we first have to take a step back and show them how to analyze technology and its role in society. We hope that with this knowledge they will not only consider policy prescriptions post hoc to remedy some of these problems, but think deeper to some of the underlying assumptions that go into data mining and the design of these algorithms. Any misconceptions they have about technology in general would be applied to their analysis of the problems that arise from data mining. To avoid that, and to make sure all students are looking at data mining from the same perspective, we believe it is important to introduce the concept of technology as a socio-technical system before moving on to the resulting issues.
For information about bias in data mining please click here.