Google presented LearnLM, a generative AI model designed for educational purposes, at the Google I/O 2024 developer conference.
Google said that the research department at artificial intelligence company DeepMind worked with Google Research to develop the LearnLM model. These models are mainly based on the famous Gemini model and aim to provide innovative and personalized learning experiences.
LearnLM features include the ability to teach students through conversations on various topics, which Google is currently looking to integrate into its various apps such as YouTube, Gemini apps, search apps, and Classroom.
Speaking at the conference, James Manyika, Director of Research, Technology & Society at Google Research, explained that the LearnLM model is designed to transform the learning experience into an engaging and personal one.
Google said it is working with teachers in a pilot program in Google Classroom to improve and simplify the process of preparing lessons and provide educational ideas, activities and materials tailored to the needs of students in the classroom.
Additionally, LearnLM Models supports Circle to Search on Android devices to solve basic math and physics problems, making it easier to ask questions and get answers when watching educational videos on YouTube.
The LearnLM module allows users to create personalized chatbots within the Gemini app that provide learning tips and training activities, tailored to each learner's preferences. For example, users can create custom bots with educational roles.
Google also plans to expand LearnLM beyond its own products by collaborating with educational organizations such as Columbia Teachers College, Arizona State University, NYU Tisch, and Khan Academy.
However, the LearnLM model may currently have some problems, such as: b. Difficulty in dealing with students' correct answers, the illusion of control, and the information provided that may be misleading.