AI guardrails advance

Lawmakers are moving to limit certain AI roles—therapy-chatbot bans are advancing in states like Maine and measures are progressing in Missouri—while commercial AI tutoring and always-on tutoring services continue to proliferate. That legal momentum and market growth together signal a policy shift: states are distinguishing between acceptable AI assistance and roles that should remain human-led. (transparencycoalition.ai) (openpr.com) (theatlantic.com)

A state can now say two opposite things about artificial intelligence at once: “do not let it act like a therapist” and “go ahead and sell it as a tutor.” Maine sent a bill on April 10 that would bar artificial intelligence from providing therapy and independent treatment recommendations, while tutoring products keep advertising round-the-clock help for homework and test prep. (mainelegislature.org) (maine.gov) (tutel.app) Missouri is moving in the same direction on mental-health use, but through a large health-care package instead of a stand-alone artificial intelligence bill. Transparency Coalition said on April 10 that Missouri is advancing a similar therapy-chatbot ban through an omnibus health-care bill. (transparencycoalition.ai) (citizenportal.ai) Maine’s bill does not ban every use of artificial intelligence in counseling offices. The measure allows licensed professionals to use it for administrative support or supplementary support, but blocks it from delivering therapy on its own or making treatment plans without a licensed clinician. (trackbill.com) (maine.gov) That line is becoming a pattern, not a one-off. The Future of Privacy Forum said last month it is tracking 98 chatbot-specific bills across 34 states and three federal proposals, with mental-health chatbots and “companion” bots emerging as two of the hottest targets. (fpf.org 1) (fpf.org 2) The concern is not just that a bot might give bad advice once. Lawmakers are focusing on systems built to simulate relationships, because a program that talks like a confidant can blur the line between customer service, friendship, and clinical care. (fpf.org 1) (fpf.org 2) Education is getting treated differently. Tutel, a German tutoring service marketed to students in grades 5 through 13, says it offers 51 subjects, free access at the entry tier, premium service at €19.99 a month, and availability “24/7,” including late-night exam prep. (tutel.app) That pitch is the opposite of a therapist’s office. A tutor can be framed as a study tool that explains quadratic equations step by step, while a therapy bot is being framed by states like Maine as something that should not make clinical judgments or act without a licensed human in the loop. (tutel.app) (maine.gov) The market keeps pushing on the tutoring side because the economics are obvious. Tutel compares traditional tutoring at about €30 an hour and roughly €240 a month with an artificial-intelligence product that says it can answer unlimited questions at any hour for less than €20 a month. (tutel.app) Researchers and education writers are also describing a shift from simple chatbots to “agents” that stay with a student across tasks and adjust help over time. Recent education coverage and research reviews describe systems that monitor quiz responses, time on tasks, and patterns of struggle to offer real-time guidance instead of one-off answers. (arxiv.org) (disco.co) (theatlantic.com) So the new rule taking shape is not “artificial intelligence yes” or “artificial intelligence no.” It is closer to: use it as a calculator, coach, or assistant if a human remains responsible, but do not let it occupy jobs where people expect judgment, duty of care, and a real professional relationship. (maine.gov) (fpf.org) (transparencycoalition.ai)

Get your own daily briefing

Scout delivers personalized news, insights, and conversations tailored to your role and industry.

Download on the App Store

Shared from Scout - Be the smartest in the room.