零三教育0—3 EDU NOTES关于

全部动态

0—6 岁教育与 AI 信息流 · 按原始发布日期排序

2

2026年7月31日周五

1
01
Springer · AI Education儿童安全同行评审儿童/教育通用73

Da Mosang Tarsier Series: Developing Digital Bullying Prevention Materials for Filipino Parents of Grade School Students

该研究开发了面向菲律宾小学生家长的‘Da Mosang Tarsier系列’数字欺凌预防材料,通过教育干预提升家长对数字欺凌的认知、态度和应对效能。

判断
将数字欺凌预防的责任延伸至家庭,填补了菲律宾本土化家长教育资源的空白,为家庭参与儿童网络安全提供了循证工具。
实践
教育机构和社区工作者可参考该材料,开展针对小学生家长的亲子数字安全教育工作坊,并评估其在本地的适用性。

2026年7月20日周一

1
01
OpenAI Blog儿童安全实践案例儿童/教育通用81

Safety and alignment in an era of long-horizon models

OpenAI discusses safety and alignment challenges arising from long-horizon AI models, which operate over extended periods. The company shares insights from iterative deployment, including new safety risks, observed failures, and improved safeguards.

判断
Long-horizon models could transform child interaction with AI through persistent agents, but they introduce novel risks (e.g., cumulative errors, manipulation over time) that are critical for early childhood contexts where sustained AI engagement may occur.
实践
Educators and developers should monitor ongoing deployment findings from frontier AI labs and advocate for child-specific safety evaluations in long-horizon models before allowing their use in educational or caregiving environments.