Data and SocietyIntroduction
Data Science is an exciting area that has helped us automate many processes using machine learning and other techniques. The impact of data science on our daily lives, and society more generally, has been tremendous. Much of data science has focused on extending and optimizing the underlying algorithms and technologies, often without considering how these algorithms and their concrete applications affect people and society. For instance, deep learning algorithms, once they have been trained, are essentially black-box routines, where nobody can explain or justify the outcomes of running the algorithm on new data: this becomes an issue when such algorithms are used for loan decisions, recommendations for parole or bail, or for hiring and salary recommendations. There are also legal issues since algorithms may discriminate against groups of individuals. For these reasons, and many others, it becomes increasingly important to think about biases, transparency, and fairness of machine-learning algorithms and training data sets. It is also important to communicate the limitations of data science algorithms in terms of fairness and accountability to non-experts who may not have the expertise to know what these limitations are.
The Data & Society course that you will take in the spring semester has two related goals. First, we will explore how data science impacts society more generally, and how to understand those impacts in our work as data scientists. Second, because data scientists have to communicate about their work to a general audience in their jobs, we will use this course as a way to improve our skills as data science communicators, primarily in writing. To prepare for the course, you'll do some reading and writing now, so that some of the issues we'll address in the spring are in your minds as you learn data science techniques in the fall. You will also get some feedback now on your writing, so that you know what will be expected in this graduate program.
To prepare for the course, you will do some reading and writing now, so that some of the issues we will address in the spring are in your minds as you learn data-science techniques in the fall. You will also get some feedback now on your writing, so that you know what will be expected throughout this graduate program.
Some commonly referenced examples of Data Science going wrong in the public sphere include the COMPAS system as profiled by ProPublica and Amazon's scrapped AI hiring tool. We will explore these and others in the class, but before we start, please list 1-3 examples of specific data-science applications that have impacted individuals or society in negative ways: these can be examples you imagine might happen or examples you have heard of.
there is no right answer here, we want you to think about these and have examples in mind as the program proceeds
In the course we will see many ways in which data science can affect society and learn strategies to mitigate potentially negative impact and conduct more responsible data science.
Which of the following topics will be covered in this class