AI tutoring can mislead

Artificial intelligence tutors can sound certain while giving students wrong, incomplete, or off-target help, and schools are starting to treat that as a classroom risk. (northernvirginiamag.com) In Northern Virginia, Fairfax County Public Schools and Prince William County Schools were early testers of OpenAI’s ChatGPT for Teachers in November 2025, but the reported use is centered on staff work such as lesson planning, translation, and adapting materials. Fairfax County Public Schools Superintendent Michelle Reid told Northern Virginia Magazine the district is taking a “measured approach” as the tools keep changing. (northernvirginiamag.com) Fairfax technology chief Gautam Sethi said district-approved systems come with enterprise privacy terms, including limits on using school data to train models. Loudoun County Public Schools has also limited faculty and staff use to district-approved programs, according to the same report published April 13, 2026. (northernvirginiamag.com) The caution comes as student use keeps rising. RAND reported in March 2026 that the share of students in middle school and above who said they used artificial intelligence for homework rose from 48 percent in May 2025 to 62 percent in December 2025. (rand.org) RAND also found that 67 percent of those students agreed that more artificial intelligence use for schoolwork would harm critical thinking, up by more than 10 percentage points in 10 months. That leaves schools trying to manage a tool students already use, even as many students doubt it is helping them think better. (rand.org) Teachers in Kentucky are responding by teaching students how to question the tool instead of treating it as an answer machine. LEX 18 reported on April 13, 2026 that Model Laboratory Schools teacher Angela Hardin was selected for the National Council of Teachers of English’s 2026 English Language Arts Artificial Intelligence Framework Project, a 50-teacher group writing classroom guidance. (lex18.com) Hardin told LEX 18 that students are using artificial intelligence “whether I like it or not,” and her class now teaches idea generation, paraphrasing, summarizing, and plagiarism boundaries. Students interviewed by the station described a shift from blanket bans toward supervised uses such as brainstorming interview questions for a podcast assignment. (lex18.com) Outside school systems, advice to students is still uneven. A sponsored practical guide published by NerdBot on April 7, 2026 said many students use artificial intelligence without teacher instruction, sometimes copying answers or using unverified information, and urged using the tools to support reasoning rather than replace it. (nerdbot.com) Federal and international guidance has been moving in the same direction. The United States Department of Education released a toolkit in October 2024 for “safe, ethical, and equitable” artificial intelligence integration, and UNESCO published global guidance in 2023 calling for human-centered rules, teacher training, and protections around privacy, bias, and age-appropriate use. (eric.ed.gov) (unesco.org) The immediate fight in schools is no longer whether students will encounter artificial intelligence. It is whether the adult in the room can show them when the machine is useful, when it is wrong, and when they need to do the thinking themselves. (northernvirginiamag.com) (lex18.com)