Recent analysis by Securly, a school safety and internet filtering firm, reveals that approximately 20% of student interactions with generative AI involve inappropriate behaviors like cheating, bullying, or self-harm. With data from over 1.2 million interactions across 1,300 districts between December 2025 and February 2026, Securly found that 2% of these interactions raised serious red flags, prompting concerns about violence or cyberbullying. Most AI usage, however, aligns with district policies, with 80% of conversations deemed appropriate when guidelines are established. Interestingly, ChatGPT emerges as the most popular AI tool among students, accounting for 42% of AI interactions. Educators are urged to maintain oversight of AI use in classrooms, focusing not only on filtering but also fostering valuable educational experiences. This data highlights the need for balanced AI implementation in schools to support safe and effective learning environments.
Source link
