������

In the Technology Sector Finnish project, we have piloted ways to incorporate the Finnish language into English-language instruction. In our first pilot conducted in technology courses, we developed Finnish-language materials for the course ELEC-C7222, Embedded Programming with Communication Devices offered at the School of Electrical Engineering. The starting point for the pilot was to summarize topics, introducing the course content in plain Finnish, and to create short comprehension exercises that would help students learn the language of their discipline in Finnish.

We presented the model to instructors in the Bachelor of Science and Technology program, and teaching assistant Rainer Ruuskanen further developed the idea for the course ELEC-C7420 Basic Principles in Networking. In the course, students explored reference models for various communication networks and the protocols at different levels of these models. The course also covered the basics of securing computer networks, such as the methods we can use to encrypt messages we send and receive over the internet.

Rainer developed short introductory videos on the course topics using AI. I interviewed Rainer about the development of the materials and asked him to share his tips to inspire other instructors.

1. Which AI tool did you use, and what did you develop with it? What is the added value of creating these Finnish-language materials?

   I used Google Notebook LLM and Gemini to create short introductory videos on various course topics in Finnish. There is a limited supply of Finnish-language instructional videos on technology compared to the availability of English-language videos. You can input English-language instructional materials into the Notebook LLM tool, and the tool can use these materials to create instructional videos in various languages, such as Finnish. Once the video is created, you can add subtitles in both Finnish and English using the Gemini tool. Anyone can create these videos using the Notebook LLM tool.

   Videos created for the course can be used for many purposes. They can be used to review topics covered in class after a lecture or to practice for an exam. The videos can also be used to learn Finnish vocabulary and improve listening comprehension. Although English is the lingua franca of the tech industry, I believe students will benefit from learning field-specific terms in the local language.

2. What challenges did you face, or where do you think there is room for improvement?

   Creating the videos was a fairly painless process, but it was challenging to edit a specific part afterward without changing other parts. When making a video with Notebook LLM, you can prompt the model to get exactly the part of the video you want, but even this kind of prompting doesn’t always work. The playback speed of the videos was also a bit too fast, so I always had to slow it down a little afterward. (By the way, “prompt” translates to “syöttötieto” in Finnish, but I don’t like this translation.)

   Many of the students were first-year students who had taken only one Finnish language course before this one. Because of this, the vocabulary in the videos was too challenging, but it was difficult to modify the vocabulary using prompts. Perhaps a possible solution would be to provide the model with materials that use simpler vocabulary. That way, the model would use the same simpler vocabulary when generating the video.

3. What further ideas appeared as a result of the pilot?

   It might be useful to try out the new GPT Real Time model, which OpenAI recently released. With the GPT Real Time model, students can communicate by speaking, and the model responds in speech (speech-to-speech). This could be a great model for students who want to learn Finnish. You could feed the model course materials, and it could discuss the course topics with the student in Finnish. It can be difficult for students to practice Finnish with native speakers, since we tend to switch the conversation to English pretty quickly. Language models could help students practice their conversational skills.

4. What kind of feedback did you receive from students?

   I received verbal feedback that the videos were too challenging for beginners learning Finnish, but the English-Finnish subtitles helped. Even though I slowed down the videos slightly, for some, the speech in the videos was still too fast for them to make out the words.

   Despite these challenges, I believe that with the help of AI, we can create more teaching materials in Finnish for students. I’ve only talked about videos here, but with Notebook, you can create things like slideshows, quizzes, and flashcards.