Speaking and Thinking for Ourselves

Jeremy Proulx
Eastern Michigan University

This assignment uses the short essay by Immanual Kant, “An Answer to the Question: What is Enlightenment?” to frame a discussion of the ethics of generative AI. In the essay, Kant defines Enlightenment as the emergence from a kind of self-imposed inability to speak for ourselves. Kant argues that pursuing the values of Enlightenment involves participating in public discourse, what Kant calls the public use of reason. These themes are explicitly couched in the context of using generative AI to help with completing assignments. What does the appropriate use of Generative AI look like? How can we utilize these powerful tools and not abdicate our moral responsibility to participate in public discourse in a meaningful and productive way? How can we use Generative AI as a writing tool and still speak for ourselves?

Learning Goals

Introduce students to the concept of Enlightenment that defined the intellectual landscape in 18^th century Europe.
Practice using AI in a class where that use is perfectly acceptable, and where we are free to talk about its strengths and limitations.
Provide a conceptual framework to be able to see past AI as merely a tool for cheating. Everyone is going to use it; how do we do so properly?
Understand that old dead philosophers still have something to offer, and that philosophy can be a valuable thing for people to choose to study in university.

Original Assignment Context

The assignment was originally somewhat ad hoc in that it emerged as just a part of a single unit on Kant and modern philosophy more generally. The assignment took place over two class periods, with an introduction to the essay in the first period and then an in-class practice session using generative AI to create essays about Kant’s essay. Over the entire term, students are responsible for submitting 15 small reflection assignments at the end of class periods, reflecting on the topic of discussion for the day. This assignment merely utilized that already-existing assignment structure, providing a little more guidance than is usual.

Materials Needed

In the first class period, students don’t need anything more than what they normally have for taking notes. During the second class period, a laptop is required, though a smartphone would work. It’s not necessary that every student has this since I break them into small groups of 2-4. As an alternative, instructors could assign students to generate some responses outside of class and come to class prepared to discuss and work with a small group to compare experiences.

Time Frame

In the context of my class, the time frame was 1 week. We have two classes/week and I present the material and the assignment on the first day and then complete the assignment on the second. Each class period is 75 minutes, leaving plenty of time for introduction, in-class work, and discussion.

Overview

This assignment began in the Fall semester of 2023 in an Introduction to Philosophy class. The class reads a short essay by the 18^th century German philosopher Immanual Kant, What is Enlightenment? In the essay, Kant defines Enlightenment as an emergence from a kind of self-imposed inability to speak for ourselves. Kant argues that pursuing the values of Enlightenment involves participating in public discourse, what Kant calls the public use of reason.

I explicitly couched these themes in the context of using generative AI to help with completing assignments. What does the appropriate use of Generative AI look like? How can we utilize these powerful tools and not abdicate our moral responsibility to participate in public discourse in a meaningful and productive way? How can we use Generative AI as a writing tool and still speak for ourselves?

The assignment takes place over two class periods. In the first, I introduce Kant’s essay and the connection it has to the use of generative AI. I also review some examples of AI generated responses to essay prompts about Kant’s Enlightenment essay and lead a discussion about their strengths and weaknesses. At the end of class, I ask students to come prepared to the next class period to experiment with Generative AI. Most students already bring laptops and those who do not are placed in a group with people who do. The second class period is devoted to experimentation with various Generative AI tools. I used ChatGPT and Bard, but any would work. I provide a few prompts based on Kant’s Enlightenment essay, students take about 20 minutes to generate responses and then each group reports back on their experience. I provide some questions to lead discussion.

Assignment

Day 1 (75 min class period)

(20 min) A short lecture introduces Kant’s Enlightenment essay and its historical context. A reading by David Hume precedes this topic and the introduction makes a connection between Kant and Hume in the context of the Enlightenment. Kant’s definition of enlightenment as an emergence from self-imposed immaturity (selbstverschuldeten Unmündigkeit) is discussed and the meaning of the German word Unmündigkeit is highlighted as significant for Kant’s idea of enlightenment. Unmündigkeit does indeed mean immaturity, but it also has an explicit connection to the German word for mouth, der Mund, and so the immaturity Kant is talking about is a kind of mouthlessness, or an inability to speak for oneself. And so the way Kant is defining enlightenment has to do with creating a society in which everyone is free to speak for themselves and where everyone takes it upon themselves to participate in public discourse in a responsible and collaborative manner.
(15 min) The lecture also introduces and highlights Kant’s distinction between the public and private use of reason. Kant defines the public use of reason as the use we make of our own critical and intellectual capacities (what Kant is here calling ‘reason’) in the public sphere, in dialogue with others. Kant defines the private use of reason as the obligation we have to use our reason in the service of external authority. Kant uses the example of a military officer having an obligation to follow orders and not question authority. He balances this need to maintain order with the additional need for citizens to be able to question authority in their capacity as citizens, in the public sphere. Students are asked to come up with some their own examples of the private v. the public use of reason.
(10 min) The generative AI aspect of the lesson is introduced with some examples of AI generated responses to the question of how Kant distinguishes between the private and the public use of reason. Significantly, ChatGPT gets this distinction wrong, almost completely backwards in fact, which helps make a point about the perils of relying on chatbots to do our thinking for us.
(15-20 min) The lesson continues with a discussion of the ethics of using chatbots to help us do our work for us. Themes from Kant’s essay, like the responsibility of thinking for ourselves and doing so in a publicy responsible manner are discussed in the context of progressively improving our society. Can one rely on a chatbot and still fulfill Kant’s requirement that we think and speak for ourselves? How so? Is there any significant difference between using generative AI and using material written by someone else? Is using generative AI a form of plagiarism? Instructors should keep track of the themes that come up in the discussion to be able to review things in the next class period.
During the last few minutes of class, students are provided with a few prompts to take home and use to generate response from an AI of their choice. Prompts are provided, but I encourage students to experiment a little, play with the prompts in an effort to generate desired results from the AI. Students are also encouraged to keep track of what they did and their experience. Students are reminded to bring a laptop or smartphone to the next class. Most already do anyway.

Day 2 (75 minute class period)

(10-15 min) A short review of the themes from Kant’s essay and especially of the issues that came up in class discussion. Students are encouraged to raise any further questions, ideas, issues, etc.
(25 min) Students are provided with a set of essay prompts to use to generate short essays with AI bots. Students work in groups of 2-4.
(30 min) Each group is asked to report back on their experience using the AI, guided by questions like the following. Which AI did you use? Did different AIs generate substantially different results? Would you feel comfortable turning these results in as your own work? Why? Why not? Do you think you could use this to build your own essay? What would you need to do? Can you use AI tools like these to complete assignments like these while still fulfilling Kant’s requirement that we speak and think for ourselves. Is the use of generative AI consistent with Kant’s concept of life in an enlightened age?
Each student is asked to submit a short reflection on their experience working with an AI.