Responsible AI Use.

1. Our Philosophy

Ethos A.I is being built around a small set of ideas: human control, local ownership, privacy-conscious design, responsible use, transparency, and honest verification of important outputs. This page explains the “responsible use” part in plain language. It is a companion to the more formal Terms of Use, Acceptable Use Policy, and AI Disclaimer.

2. Human Review Matters

AI output can be incorrect, incomplete, outdated, biased, misleading, or inappropriate for a specific situation. That is true for AI generally and it is true for Ethos A.I. The reasonable default is to treat AI-generated content as a helpful starting point that a human then reviews, verifies, edits, and takes responsibility for — especially for anything consequential.

Use Ethos A.I to think faster, not to skip thinking.

3. High-Stakes Decisions

Ethos A.I should not be used as the sole decision-maker for high-stakes decisions, including:

• medical diagnosis or treatment;
• legal advice;
• financial or investment decisions;
• employment decisions;
• credit decisions;
• housing decisions;
• education admissions;
• insurance eligibility;
• emergency response;
• safety-critical systems.

These decisions require qualified human review and, in many cases, specific regulated safeguards. AI assistance is not a substitute for qualified professional judgment. This does not mean Ethos A.I can never help in these fields — it means the human, and the appropriate qualified process, must remain in charge.

4. What Ethos A.I Is

Ethos A.I is being designed as an assistant that can help you think, draft, summarize, organize, and search — on hardware you control, using material you have approved. It is intended to make private local AI easier to use.

5. What Ethos A.I Is Not

Ethos A.I is not:

• a replacement for licensed professionals or trained specialists;
• an oracle of guaranteed truth;
• an autonomous system authorized to take consequential actions without human approval;
• a substitute for your own judgment;
• a substitute for the security, privacy, and backup practices you already use.

6. Transparency

We aim to describe Ethos A.I honestly. Development previews may differ from final functionality. Planned features may change before release. When something is speculative or uncertain, we say so. We do not use language like “100% private,” “unhackable,” or “impossible to access.” See the AI Disclaimer and Privacy Policy for the details.

7. Design Intentions

Ethos A.I is being designed around:

• Local-first operation — core supported features are intended to run on the user's hardware;
• Approved-folder access — the software is designed to work only with folders the user has explicitly authorized, not to automatically scan the entire computer;
• User-controlled knowledge — you decide what is indexed, what is stored, and what is deleted;
• Documented external services — if an optional feature involves network connectivity, it should say so and require the user to enable it.

These are design intentions, not guarantees. Final documentation for supported hardware, connected services, and behavior will be published before the software ships.