1 in 5 high schoolers had a romantic AI relationship, or knows someone who has (www.npr.org)

A new national survey from the Center for Democracy and Technology (CDT) finds AI has become deeply embedded in U.S. schools—and with it new social and safety risks. Among roughly 1,000 high‑school students, 800 sixth–12th grade teachers and 1,000 parents surveyed, 86% of students, 85% of educators and 75% of parents reported using AI in the last school year. Nearly 1 in 5 high schoolers said they or someone they know had a romantic relationship with an AI, and 42% said they or someone they know used AI for companionship. Students also reported using school devices or software for personal AI conversations (31%). CDT notes that higher levels of school AI use—defined in the study as 7–10 teacher use cases and 4–6 student use cases—correlate with more exposure to data breaches, AI failures, and AI‑generated deepfakes (manipulated images/videos) used to harass or bully students. Technically and practically, the report flags several implications: frequent school use of AI aligns with a higher reported rate of large‑scale data breaches (28% vs. 18% for lower‑use teachers), more system malfunctions, and erosion of community trust when monitoring tools produce false alarms or lead to punitive actions. Only 11% of teachers received training to respond to harmful student AI use, underscoring gaps in AI literacy, privacy safeguards and policy. CDT’s findings argue for stronger data‑protection practices, targeted educator training, and student-centered safeguards to balance AI’s instructional benefits with risks to wellbeing, privacy and equity.