Limitations and what’s next
This is a first step in a longer journey to improve learning in ChatGPT. Today, study mode is powered by custom system instructions. We chose this approach because it lets us quickly learn from real student feedback and improve the experience—even if it results in some inconsistent behavior and mistakes across conversations. We plan on training this behavior directly into our main models once we’ve learned what works best through iteration and student feedback.
We’re also exploring functionality to make study mode more engaging and helpful for students including:
- Clearer visualizations for complex or text-heavy concepts
- Goal setting and progress tracking across conversations
- Deeper personalization tailored to each student’s skill level and goals
The research landscape is still taking shape on the best ways to apply AI in education. OpenAI is enabling further research on learning and AI through various partners in its NextGenAI initiative, and working with experts from the SCALE Initiative at Stanford University’s Accelerator for Learning(opens in a new window) to study and share how AI tools influence learning outcomes.
As we run longer-term studies on how students learn best with AI, we intend to publish a deeper analysis of what we’ve learned about the links between model design and cognition, shape future product experiences based on these insights, and work side by side with the broader education ecosystem to ensure AI benefits learners worldwide.
