Language Learning Practice with GenAI Simulations

Carolin Fuchs
Course Subject:Undergraduate German
Student Level:Beginning 1 (first-semester) and Intermediate 2 (fourth-semester)
Number of Students:15 students per section
Developed by:Carolin Fuchs, Teaching Professor, Department of World Languages and Cultures

What Students Did

During the last four weeks of the semester, students leveraged Claude’s chat feature to engage in text-based simulated conversations related to authentic tasks such as finding an apartment or applying for a job. They analyzed and compared their chat transcripts with their peers, and reflected on their experiences.

Purpose

The purpose was to help students improve their skills and confidence in using German in everyday conversational scenarios such as interviewing for a job or searching for an apartment. The simulations were grounded in the Language Learning Beyond the Classroom (LLBC) framework for language learning, which is an informal learning approach to learning a foreign language outside of formal structures (Benson, 2017). In addition to conversational fluency, the simulation supported student agency and self-directed learning, including their capacity to use AI purposefully in developing proficiencies that are best learned over time through practice.

Assessment

The conversation simulations in both the beginner and intermediate classes were formative assessments that contributed to the student’s engagement grade as well as the overall course grade. To receive credit, students were required to submit their chat conversation transcripts along with reflections about the process.

Chat transcripts were analyzed for fluency, pragmatic learning, and grammar improvements. Reflections were analyzed for growth in confidence and language development over time. The analyses showed that students improved in both confidence and competency. Their more passive and simplistic responses in the first simulation gave way to more complex conversations, initiation of topics, and thinking more directly in German rather than translating.

Faculty Reflections

Students appreciated real-world conversational contexts and the error correction the simulation provided. Students expressed a desire for longer simulations as well as voice chat functionality. For more advanced levels of language learning, more scaffolding of domain-specific vocabulary practice is needed. In addition, students may need more guidance on how to comfortably end conversations with the AI.


Step-by-Step Directions

Step 1Assign the Simulation Task, including the information students need to run the simulation independently outside of class:

  • Situation (e.g., job search)
  • Objective (e.g., practice the content and vocabulary associated with the situation)
  • AI role and Student role information (e.g., the bot is the employer, you are the job applicant)
  • Tips (e.g., ask only one question at a time, ask the bot to repeat something if you don’t understand)
  • Tasks to do prior to the simulation (e.g., review vocabulary, identify a job in your field using a job portal, review 10 most commonly posed interview questions)
  • Prompts to use during the simulation

NOTE: Simulation task details, including prompts, are provided in the related materials.

Step 2Individually, outside of class, students engage in the simulation in German. When they are finished, they create a chat log link to share with the instructor and their peers.
Step 3After the simulation, students reflect on their experience as a group. They respond to several prompts in English within a discussion in the learning management system (Canvas), including the link to their simulation chat. Example guiding questions include:

  • What went well in this assignment? Were there any differences compared to the first simulation?
  • What surprised or puzzled you?
  • Were there any new words you learned?
  • Was there anything that didn’t make sense?
  • What would you like to do differently next time?
Step 4Review the discussion and logs to determine strengths and gaps in student interactions with the bot, reported gains in confidence, and any other reflections about the experience that could be useful for further work with students individually or as a group.
IterationStudents repeat the work as follows:

  • Beginner level students – engage in four simulations (Apartment search, Meeting on campus, Making plans, Professions and dream job).
  • Intermediate level students – engage in two simulations (Job application, Professions and dream job).

Related Materials

Download PDF version of this assignment

Return to AI Gallery Home