Campfire Session — World Languages

Introduction • 00:00

• Lulu Gao introduces the session. The focus is on using Flint in World Language classrooms.

Ice-breaking News • 02:24

• A discussion is initiated about AI's impact on languages. Concerns are raised regarding language extinction due to dominant AI models.

• Carolin Escobar discusses the cognitive benefits of learning languages.

• Karina DeMarco highlights AI's role in facilitating access to lesser-known languages.

Student experience of language activities in Flint • 07:24

• Lulu Gao presents a language activity example in Flint. The platform's individualized practice and feedback for students are discussed.

• A question about session limits is raised. Students can create unlimited sessions for practice.

• Concerns about question recycling of questions during practice sessions are discussed. Students may encounter the same questions unless instructed otherwise.

• Feedback on the pronunciation capabilities of Flint is provided. Suggestions for improving tone correction in Mandarin Chinese are made.

• Discussion on how Flint handles different accents. It was noted that Flint understands various accents and dialects well.

Creating a language activity in Flint • 19:30

• Demonstration of creating activities using Flint's platform. The process of customizing activities based on student proficiency and topics was explained.

• A question about activity setup is raised. The setup includes writing-only or speaking-only activities, as well as activities that can do both.

• Memory limitations during activity creation are discussed. A 10x increase in memory has been achieved within the last year and the technology is expected to continue improving

• Feedback capabilities of Flint are discussed. Feedback can be given in different languages.

• One teacher shares experiences with Flint's feedback system. The issue of students not seeing teacher feedback is highlighted. This may have been students visiting the wrong link, but Lulu agrees notification of students can be improved here.

Examples of activity use in language courses • 41:00

• A variety of templates for activities is presented. Examples include a train journey and an image analysis of a protest in France:

  • Conversational roleplays

  • Listening comprehension

  • Reading comprehension

  • Describe a picture

  • Vocabulary practice

  • Oral presentations

  • Quiz on content

• Suggestions for when Flint activities can be used are also shared:

  • Bellringer/warm-up

  • New concept introduction

  • Project feedback/research

  • Individualized discussion

  • Review for quiz/exam/presentation

  • Independent study/tutoring

Tips for activity set-up • 44:17

• Tips for using Flint effectively are shared. Suggestions include setting up language preferences and communication methods.

• Strategies for adjusting AI behavior and feedback are discussed. Specific prompts can be used to guide student interactions with Flint.

• The complexity of feedback is discussed. It is noted that feedback depends on the rubric used.

Teacher shareout and more Q&A • 49:55

• The resources Flint uses to set language levels are explained. The CEFR and ACTFL systems are mentioned as guidelines.

• A specific example of using the rubric in practice is shared. Adjustments to AI instructions for student engagement are highlighted.

• The request for HSK levels is discussed. It is suggested to reference HSK levels in activity creation.

• Tristan Ming shares innovative uses of Flint. A chatbot was created for students to engage with characters from picture books.

• Lulu Gao addresses a question about saving activities. It is clarified that activities can be saved but chat history will be lost.

Conclusion • 58:05

• A session on crafting AI policies is announced. Discussions on various schools' progress are shared.

• Teachers are encouraged to connect via a Facebook group. The group is active with teachers sharing ideas and advice.

• A suggestion is made to test activities from a student's perspective. Emphasis is placed on setting expectations about AI's limitations.