Policy
New State Laws on AI Companions Create Compliance Puzzle for Schools
Regulations targeting chatbots with emotional features could restrict tutoring tools, forcing districts to balance personalized learning against legal risk.
Key takeaways
- Eighteen state laws now regulate AI companion chatbots, with regulations distinguishing them from standard chatbots based on emotional engagement features.
- Adam's Law, effective July 2027, will require one-hour session limits and disabled conversational memory by default for children, with no express K-12 exemption.
- Educational chatbots designed for tutoring could face restrictions if they sustain ongoing personal conversations with students across multiple interactions.
- Districts face uncertainty about whether they or vendors qualify as the legal "operator" when providing students access to third-party chatbot tools.
- School leaders must now weigh the educational value of personalized AI tutoring against legal compliance requirements and student safety risks.
School technology leaders face a new regulatory challenge as states enact laws targeting AI companion chatbots—restrictions that could apply to educational tools designed to tutor students or provide academic support.
The legal landscape shifted rapidly in 2025 and 2026, with 18 state laws now on the books addressing chatbot use, according to privacy expert Amelia Vance, president of the Public Interest Privacy Center. During a session at the California IT in Education conference last week, Vance outlined how these regulations distinguish AI companions from standard chatbots based on their capacity for emotional engagement.
"It's a system capable of meeting a user's social needs that can sustain a relationship across multiple interactions," Vance explained, noting that even tutoring tools could face restrictions if designed to maintain ongoing personal conversations with students.
The compliance challenge for districts
The question of which educational tools qualify as AI companions remains murky. Chatbots that exhibit humanlike characteristics—including expressions of affection, encouragement, or emotional connection—may trigger regulatory requirements, even when their primary purpose is academic.
One upcoming law, Adam's Law, illustrates the complexity. Set to take effect July 1, 2027, the legislation includes exemptions for higher education and workplace use but does not expressly exempt K-12 schools. The law requires covered operators to disable persistent conversational memory by default for children, limit sessions to one hour, and cap daily use at two hours. Only parents or guardians—not school employees—can modify these default settings.
Vance noted uncertainty about whether districts or vendors would be considered the "operator" when schools provide students access to third-party chatbot tools.
The risks behind the regulations
The legislative push follows troubling incidents involving young users and conversational AI. Vance recounted the case of Adam Raine, a 16-year-old who initially used ChatGPT for schoolwork while away from school for medical reasons. According to a New York Times investigation, Raine's conversations with the chatbot eventually turned to suicidal thoughts. The AI sometimes encouraged him to seek help but in other exchanges provided dangerous information or discouraged him from confiding in others. Raine later died by suicide.
"The longer the conversations go on, the more the guardrails break down," Vance told attendees. "But the longer the conversation goes on, the more you can personalize tutoring, provide mental health support, make a kid feel seen."
This tension between personalization and safety sits at the heart of the compliance dilemma facing schools.
Practical considerations for school leaders
Conference attendees raised practical concerns about implementation. One district leader noted that her schools allow staff and ninth-grade students to use Google's Gemini chatbot, questioning whether students would be better served learning to use these tools under educator supervision rather than experimenting independently at home.
Another attendee wondered whether monitoring students' chatbot conversations could backfire, prompting them to abandon district-provided tools in favor of personal accounts that lack school oversight.
Vance suggested that districts reconsider how much conversational memory chatbots truly need to retain. "It doesn't need and really shouldn't have a relationship with the child," she said, encouraging school leaders to limit conversational memory and educate students about potential risks.
Why it matters
These regulations force school technology leaders to make difficult tradeoffs between educational innovation and legal liability. As conversational AI becomes more sophisticated and embedded in learning platforms, the line between tutoring tool and AI companion will continue to blur. Districts must now evaluate not just whether a chatbot is pedagogically effective, but whether its design features could trigger compliance obligations or expose students to harm. With 37 states introducing 124 chatbot-related bills in 2026 alone, the regulatory environment will likely grow more complex before it stabilizes.
These details were first reported by Julia Gilban-Cohen for the Center for Digital Education.