AI can support a child’s progress. It also needs clear boundaries, careful checks, and adults who stay involved.
Check the question first
AI-generated math content can occasionally be incorrect. The planned live system uses a separate verification step to independently solve each generated question and check wording, answer choices, and difficulty before it reaches a student.
Verification reduces risk; it does not eliminate it. The current demo uses prewritten, mathematically checked examples.
Explain without discouraging
Feedback should be concise, supportive, and appropriate for the student’s learning level. A mistake is evidence about what to practice next, not a judgment of ability.
An answer alone cannot prove why a mistake happened. When reasoning is unavailable, a diagnosis should be presented as a possibility.
Keep adults in the loop
GBB One is intended as a learning tool, not a replacement for teachers or professional educational judgment. Parents and educators should remain involved in the student’s learning.
The planned account model puts parents in control of student profiles. There are no child accounts in this public preview.
Keep practice focused
The product is designed around structured mathematics practice, rather than an open-ended social chatbot. Planned safeguards include age-appropriate prompts, limited personal data, and review of unexpected content.
These safeguards are still being developed and evaluated before broad public availability.
STUDENT PRIVACY
Learning data deserves care.
Parent-managed profiles and data minimization are central to the planned service. Read what this website collects today and what we’re designing for launch.
If you spot an unclear question, incorrect explanation, or safety concern, please tell us. Include the problem and what seemed wrong; leave out personal student information.