Abstract
Be is a proposed organization dedicated to making safe, professionally governed artificial intelligence emotional support freely accessible to people who need someone to talk to. Be proposes an AI emotional-support system designed around a foundational distinction: an AI must never confuse what it observes with what it infers. Version 2.0 further develops the proposal's framework for adaptive conversational communication through the principle Adaptive Tone, Fixed Boundaries. Be may adapt its tone, vocabulary, pacing, humor, warmth, and conversational style to meet an individual where they are, while maintaining fixed ethical and relational boundaries. The system should allow human-like communication without creating a false human identity, adapting the interface without fabricating the relationship. The proposal incorporates Red Tape Theory™ as a proposed safety methodology: Observe. Record. Question. Do not force the axiom. Interpretation must never be represented as observation. Be does not propose to diagnose users or replace licensed mental-health professionals. Instead, it proposes a professionally governed pathway through which an AI system may recognize when AI-only support may no longer be an appropriate level of support and facilitate access to human care. A proposed charitable human-care program, If You Get It, You Get It, would seek to use charitable resources to help people access licensed professional care when professionally established safety criteria indicate that additional human support may be appropriate. Version 2.0 expands the proposal's treatment of personalization as a safety boundary, including the distinction between adapting communication to an individual and constructing a misleading simulated relationship. It also addresses clinical governance, privacy, autonomy, human handoff, crisis safety, documentation, intellectual-property considerations, research and evaluation, and a potential future child-focused system provisionally called NewBee. This publication is a founding proposal and does not represent a deployed clinical service, licensed mental-health treatment program, clinical efficacy claim, professional endorsement, or completed charitable program. Be begins with the question: Where could this harm somebody?