Keywords: AI, ChatGPT, transcription, digital recovery, humanities, humanity, 19th century

Teaching Materials Developed by: Dr. Abigayle Farrier, University of North Texas

Peer Reviewed by: Jina DuVernay

Formatted by: Alice Martin, Rutgers University

PDF version of this assignment.

Introduction:

This activity was originally designed for a 3000-level course, Academic Writing in the Humanities. I have also successfully taught it in 1000- and 2000-level American literature courses. The depth of discussion and critical analysis will vary depending upon student familiarity with this type of work, but student engagement and interest were strong in both upper- and lower-levels. This activity could also be taught by librarians as part of a lesson about primary sources, or it could work well as a collaborative/co-teaching effort between instructors and librarians.

I first taught this activity in the wake of what I thought was the great nightmare of my career: my students were using AI to write their essays, and AI was giving them information that was nowhere near the ballpark of truth. Zelda Fitzgerald’s short story, “The Iceberg,” for example, suddenly focused on the iceberg theory in psychology; Louisa May Alcott’s novel, Work, became a short article and was condensed to one chapter of woefully incorrect summary; and I was at my wit’s end. I have now taught this activity over the course of three semesters—since the mass advent of ChatGPT. It has gone through a few different iterations over the several courses I have taught it in, and the version below is one that I have found engages students and acknowledges some of AI’s benefits while also driving home the larger point that AI, in all its glory, can never replace the humanity of our work.

Student response to both the reading and the activity has been overwhelmingly positive. For my upper-level students in particular, this activity offers a glimmer of hope as they prepare to graduate and enter the workforce with their newly minted humanities degree. My lower-level students have become motivated to learn more about digital recovery, digital humanities, and the ways in which little-known writings from the 19th century can have significant impacts on our 21st century lives.

For lower-level classes, the most challenging part of the activity is the attempted transcription. This is a place where students can find themselves frustrated quickly. If your class has no experience reading cursive, I recommend giving them no more than 5 minutes to attempt transcription on their own. It is also helpful to clarify the 19th century long “S,” ampersand, and any other manuscript difficulties in the texts you choose. Sticking with the Alcott letters linked below is helpful; the context provided gives students a starting point. Even if the students are unfamiliar with cursive, they can usually get a few words from the context. This helps to minimize frustration.

Far and away the most rewarding part of the activity is the final step, in which students get to hold and look at a 19th century friendship book. They get to see the real human hair bound into the book, feel the leather binding, admire the gilt edges, and truly appreciate the love and care that went into the creation of this artifact. This moment provides the “a-ha!” breakthrough. At this point, I often share stories about other moments of great recovery work born out of humans doing what we do best: caring about the lives and stories of others. Henry Louis Gates, Jr., for example, was the only bidder on a handwritten manuscript that would later become known as the first novel written by an enslaved woman. Bernadette Farquhar found a copy of the first Antiguan novel in a trashcan outside a library and devoted part of her career to an extensive recovery effort of the author’s life and work. Even my students’ own English professor found a friendship book for sale and haggled the seller down to a price affordable to a graduate student, all because of the story—a young man solicited writings and contributions from every important person in his fiancée’s life in order to present her with a physical representation of just how loved she was. We ultimately conclude that while AI can be a legitimate and useful recovery aid, the process of recovery still relies wholly on humans and our interest in finding and preserving “lost” moments, people, and literature from the past.

When you teach this activity, if the texts below do not resonate with you, I recommend switching them out for some that do. This activity will work the same with a different author’s letters or with any handwritten materials. In my opinion, this activity works best when students can see the instructor’s passion.

Activities/Handouts/Discussion Questions:

Learning Outcomes:

  • Work collaboratively in groups
  • Define the following terms: recovery work, digital archives, digital recovery
  • Practice transcription skills by transcribing 19th century handwritten letters
  • Determine ethical uses for AI in the classroom and in digital recovery
  • Learn best practices for handling physical archival materials
  • Engage in primary source archival work

Task: In this activity, students work together to transcribe both digitized and physical handwritten materials. They explore the uses of AI in transcription and recovery work and engage broadly with the uses of AI in the humanities. This activity is an introduction to digital recovery and the digital humanities.

Purpose:The purpose of this activity is to introduce students to digital recovery and the digital humanities while exploring potential ab/uses of AI. Students will familiarize themselves with the recovery process, explore 19th century artifacts, and learn about the digitization process.

Criteria: Each student is to actively participate in the totality of our class activity for the day. Detailed instructions, which should be followed by each student, are below.

Directions:

  • Work collaboratively in groups
  • Define the following terms: recovery work, digital archives, digital recovery
  • Practice transcription skills by transcribing 19th century handwritten letters
  • Determine ethical uses for AI in the classroom and in digital recovery
  • Learn best practices for handling physical archival materials
  • Engage in primary source archival work

Grading: This is an exploratory and discussion-based activity that is not graded through formal assessment. Students are expected to actively participate, and the instructor provides oral feedback throughout each step of the process.

Contextual Materials/Resources/Further Reading: