At BuddyWithMe, our mission is simple: make focused, confidence building learning feel personal and doable for every student with ADHD (and any learner who benefits from structure). This fall, we began piloting Google’s LearnLM, a family of AI models fine-tuned for teaching and learning. Our early experiments suggest LearnLM could help us deliver clearer explanations, smarter practice, and more “stickiness” in study sessions all while staying aligned with learning-science principles.
What is LearnLM in a nutshell?
LearnLM is Google’s education-tuned set of models (built on Gemini) designed to make learning more active, personal, and engaging. It emphasizes tutoring behaviors like step-by-step reasoning, scaffolding, and Socratic prompts rather than just giving answers.
Developers (like us) can target “teaching and learning” system instructions with LearnLM via the Gemini platform, which is specifically described as aligning to learning-science principles.
What we’re piloting
- Focus Modes with “learning partner” prompts. Nudges to start, checkpoint, and reflect tuned to ADHD-friendly pacing.
- Notebook-style study packs. Turning class materials into interactive guides with formative checks and audio overviews where appropriate. (We’re watching Google’s research experiments closely here.)
Guardrails and responsibility
We’re proceeding with safety first: clear boundaries around what the AI can/can’t do, transparent prompts, and regular human review. We’re also building opt-in feedback loops so learners can flag confusing explanations or ask for more scaffolding. We’ll publish more about our evaluation approach as we collect sufficient data over the semester; Google’s recent LearnLM benchmarking work is a helpful north star for measuring real learning support, not just answer accuracy.
The bottom line
Our early exploration of Google LearnLM shows real promise for making tutoring feel more human, more structured, and more motivating especially for students who benefit from clarity and gentle momentum. We’re excited to keep sharing what works (and what we refine) as we expand our pilots.