Responsible AI & Acceptable Use Policy

Scope and accountable ownership

This policy governs the RealtrAI marketing website and the use of RealtrAI product workflows. It is owned by the President of Trunnion AI together with the Chief Technology Officer, who are accountable for its content and enforcement. The policy is reviewed quarterly and whenever a material prompt, model, provider, or workflow change ships; the effective date above changes with each revision. Trunnion AI, LLC maintains this policy as part of its alignment to the NIST AI Risk Management Framework.

How RealtrAI is designed to be used

RealtrAI provides task-specific AI drafting and calculation workflows for licensed real estate professionals. Every generated output is a draft. The intended workflow is: supply your own accurate property facts and assumptions; generate; review the draft against property facts, Fair Housing requirements, brokerage policy, MLS rules, and applicable law; then use it under your own professional judgment. Nothing auto-publishes, and the terminal product action is copying a draft you have reviewed.

Prohibited uses

You may not use RealtrAI, and RealtrAI must not be configured or prompted, to:

• Steer or express preference on protected classes. Generating, refining, or publishing any listing, advertisement, or client communication that states or implies a preference, limitation, or exclusion based on race, color, national origin, religion, sex (including sexual orientation and gender identity), familial status, disability, or any additional state or local protected class - including proxy phrasing that describes the intended buyer or renter rather than the property (for example "perfect for young families," references to houses of worship or school demographics as buyer signals, or "safe neighborhood" used as a coded descriptor).
• Target or exclude discriminatorily. Using generated content to target, exclude, or tier housing-related marketing audiences on a protected-class basis.
• Misrepresent property facts or data. Presenting generated text or calculator arithmetic as verified property facts, sourced market data, comparables, or a performance benchmark, or publishing generated claims you have not verified.
• Substitute AI for professional determinations. Presenting any output as a Fair Housing determination, appraisal, legal, tax, or investment advice, or a brokerage instruction.
• Bypass human review. Automating, republishing, or delivering generated content to clients or the public without the required professional review, or removing review steps from a team workflow.
• Act unlawfully or deceptively. Any use that is unlawful, deceptive, harassing, or that violates MLS rules, advertising regulations, another person's rights, or the applicable written agreement.

Violations may result in suspension or termination as provided in the applicable agreement, and we may refuse workflows that are designed to produce prohibited output.

The risk process, mapped to the NIST AI RMF

• GOVERN. This policy, its named accountable owners, and its quarterly review cadence. Public boundaries are maintained on the Security posture page and are not marketed beyond what is implemented.
• MAP. A public inventory of all seventeen tools with each tool's job, intended use, and limitations, on the tools pages and in the AI Transparency Notice. We are extending that inventory into an internal AI system register recording model, provider, version, data inputs, and Fair Housing exposure per tool.
• MEASURE. Automated release checks run as blocking gates in the website's build pipeline today. A Fair Housing evaluation suite for the generation workflows - adversarial prompts scored for protected-class references, preference language, and proxy terms - is a committed roadmap control; its methodology and measured rates will be published when it ships. Until then, RealtrAI does not claim, and must not be represented as providing, a verified automated compliance screen.
• MANAGE. Draft-only outputs, no auto-publishing, human review required before use, and a reporting path: report a concerning output, suspected steering language, or any AI incident to contact@viceroynm.com. We acknowledge reports within two business days, and confirmed corrections ship through the release process.

Fair Housing guidance

The dedicated Fair Housing page teaches the review method we expect every user to apply: identify preference, steering, and protected-class references; rewrite to describe the property, not the buyer; and obtain professional review under brokerage, MLS, and local requirements before publishing. That guidance is educational and is not legal advice or a compliance determination.

Questions

Questions about this policy, procurement and diligence requests, or reports of misuse may be sent to contact@viceroynm.com or mailed to Trunnion AI, LLC, 8100 Wyoming Blvd NE, Ste M4-301, Albuquerque, NM 87113, United States.