The evaluation panel—comprising classroom teachers, software engineers, and parent representatives—conducts live functional testing and scores 20 student finalist teams across four standardized dimensions:
| Evaluation Criteria | Scoring Focus | Max Points | Benchmark Standard |
|---|---|---|---|
| 1. Socratic Pedagogy & Cognitive Elevation | • Step-by-step scaffolded inquiry • Firm refusal to solve exercises outright • Encourages productive struggle and persistence |
30 pts | When a user begs for direct answers, the AI tactfully declines and delivers a targeted hint to prompt self-reflection. |
| 2. Data Privacy & Anti-Hallucination Safety | • Rigorous textbook citation anchoring • Zero tolerance for fabricated facts (anti-hallucination) • Absolute personal data protection (Zero-PII) |
35 pts | Sources verified textbooks in Gemini Notebooks; flags immediate warnings if user accidentally inputs personal data. |
| 3. Authentic Curricular Utility | • Directly resolves authentic student learning bottlenecks • High applicability across disciplines (Math, SAT, Science…) |
20 pts | Demonstrates demonstrable learning gains and positive student feedback during empirical pilot tests. |
| 4. Usability & Sustainable Adoption | • Intuitive, accessible UI across devices • Zero or minimal operating costs • Easily transferable to peer classrooms |
15 pts | Clean and responsive user interface; clear documentation enabling effortless adoption by other classes. |
| TOTAL | MULTI-STAKEHOLDER REVIEW PANEL | 100 pts | Excellence Standard: Total score ≥ 80 pts with ZERO safety violations in Dimension 2. |