What does it take to put AI into practice in schools? This brief examines what California educators learned as they designed and tested AI-enabled solutions through Silicon Schools Fund’s AI Exploration Community of Practice.
In its second year studying the community, CRPE analyzed educator-reported data from 66 pilots, observed meetings and presentations, and interviewed eight pilot leads. Their experiences suggest that early AI implementation may produce better judgment and clearer uses for the technology before measurable improvements in practice.
Key findings
- AI pilots required substantial time, labor, and resources. Educators tested tools, refined prompts, troubleshot errors, and adjusted instruction. Even when their goal was to save time, implementation required more work than many anticipated.
- Students’ concerns shaped implementation. Students questioned AI’s reliability and its effects on creativity, critical thinking, and the environment. Educators needed time for honest conversations about why they were using AI and how to evaluate its outputs.
- Structured use helped students become more discerning. Educators reported that opportunities to use and question AI helped students recognize its strengths and limitations, consider appropriate uses, and identify overreliance.
- Educators and students worried about effects on their relationships. Tasks like feedback and grading can be important opportunities to build trust and understand students. Although pilot leads did not report actual harm to these relationships, concerns about what AI might displace persisted.
- Mixed results still produced valuable learning. Every interviewed pilot lead found the experience worthwhile, including those whose pilots fell short of their goals. Educators developed a clearer understanding of AI’s capabilities, limitations, and conditions for successful implementation.
Recommendations
The brief calls on schools and districts to invest in the conditions needed to implement AI purposefully and responsibly.
- Build in structured testing and adaptation. Give educators time to test tools on authentic tasks, gather student feedback, adjust their approaches, and switch tools when necessary.
- Expect learning before returns on investment. Treat early pilots as opportunities to understand what works and what support is needed before scaling.
- Build usable evidence systems. Help educators gather practical evidence on feasibility, time savings, learning outcomes, educator and student experiences, and unintended consequences.
- Keep trust and human relationships at the center. Give educators and students a voice in decisions about AI tools and guardrails, and consider which interactions should remain teacher-led.
Educator-led pilots may not immediately transform teaching and learning. But they can build the knowledge and judgment educators and students need to make informed decisions about AI.
