Philosophy of Algorithms

A unique interdisciplinary space to discuss algorithms
In the seminar Philosophy of Algorithms (part of the Science in Perspective-program) students from computer science, engineering and natural sciences and humanities discuss philosophical questions about algorithms with a teacher from philosophy. For example, at the beginning of the term the students consider what an algorithm really is. By talking to computer scientists, engineers, philosophers, cultural studies students and discussing a philosopher’s article about what an algorithm is, the students realize that there are different implementations of algorithms in theoretical, technological and social contexts. The conflicts that the various implementations and interpretations entail are not guaranteed to dissolve in the seminar, but by actively listening to the others’ views, replying to these concerns, and sharpening their own arguments, the students develop a more thorough understanding of, e.g., a definition of algorithms. Individually, the students may still conclude that computer scientific definitions are most accurate, but they understand that other definitions are crucial in more applied, technological context.
Throughout the course the participants continually realize and reflect current debates about algorithms and contribute to philosophical and social reflections about algorithms by discussing philosophical articles on algorithms. They recreate academic and public discussions and develop their own, genuine positions in the field of discussion which they take back into their studies and their work.

Listening actively and including the others’ perspective
Students read philosophical texts involving issues with algorithms in preparation for the class. In class they discuss the texts independently in small groups and in a teacher-led setting in the plenum. The teacher is both a moderator and a participant of the class discussions. These discussions are interdisciplinary and open: Every student has space to take up and articulate their own standpoint and perspective. At the same time, all students are reminded to listen to and include the perspectives of the other participants from other disciplines or with opposing views. Some of the questions that they consider are: Can we define algorithms in a way that is precise enough for computer science and understandable for people who merely employ algorithms? Which different agents can we distinguish for algorithms? Which ethical questions are pertinent when algorithms are used? Is there a difference between algorithms in high-stakes decision making and algorithms in low-stakes decision making? What explains biases in algorithms? How should we deal with biases in algorithms? Are computer scientists responsible for the algorithms that they program?

The texts and the students’ expertise and experiences are the basis to all class discussions. The teacher reviews the current research literature and public discussions throughout the seminar so that she can suggest topical and current changes to the curriculum that are relevant to the specific discussions in class. For example, she may suggest to the students to include a session on newspaper articles, podcasts or radio reports on the effects of algorithms in administration on poverty. She is thus able to take up cues from the students and react to insights and questions in the class sessions.
In order to gain the credit points the students can choose between submitting six short essays answering a question about the previous session, writing a longer essay about, or giving a class presentation on one of the articles read in the seminar. The teacher gives immediate, individual feedback for the six short essays and the students can thus improve their writing skills and ability to express arguments and objections. The teacher presents a selection of the students’ insights from their essays at the beginning of the subsequent session so that the sessions are connected and the group learns about further observations concerning the previous topic and what are features of good and clear philosophical arguments.

In their small-group and large-group discussions the students carve out recurring topics, problems, questions and solutions in philosophy of algorithms. They relate different viewpoints to one another and to their own perspective, they uncover and question hidden assumptions and recognize conflicts of interests. To do so they must articulate their own considerations and arguments to students with other backgrounds. They have to include the other participants in deciding how to present their claims and how to illustrate and substantiate their views. Understanding the other participants’ arguments and concerns is thus a central ability for contributing to the class discussion.

Teachers learn from the students’ new perspectives
All of the above insights also apply to the teacher who has to be flexible and open for the students’ input and perspectives. The teacher’s task is to bring the students into conversations, to bridge potential misunderstandings, to show which arguments are successful, how the students can connect to other people’s perspectives, to point out implicit disagreements and agreements.
In dealing with the perspectives of the other participants the students – and the teacher – also reflect their own discipline: How does my field and my work appear to non-experts? Do I feel that my field and my views are adequately represented or not? How do external perspectives on my work affect my own work?

Result and Outlook
By including computer scientific, philosophical and social perspectives the students learn to take up and defend their viewpoint and views confidently and reflectively. They integrate social and philosophical perspectives in their understanding of their own discipline and on algorithms and thereby recognize the place of their own thought and action in science and society.