# OpenAI Brand Voice

> A clinical, utilitarian, and authoritative voice that prioritizes clarity and institutional scale over marketing flair.

## Positioning
OpenAI is an AI research and deployment company building tools for individuals, businesses, and developers. It positions itself as a foundational utility "built to benefit everyone" with a heavy emphasis on safety and institutional transparency.

## Voice principles
*   **Direct:** Uses short, declarative sentences and active verbs to describe complex technology.
*   **Institutional:** Maintains a neutral, objective tone similar to a research journal or a public utility.
*   **Functional:** Focuses on what the tool does or what the research proves, avoiding superlatives and hype.
*   **Responsible:** Explicitly categorizes content by "Safety" and "Research" to signal a methodical approach to development.

## Tone by context
| Context | Tone |
|---|---|
| Product Announcements | Descriptive and matter-of-fact. Focuses on utility and new capabilities. |
| Research/Safety Updates | Academic and precise. Uses technical terminology without apology. |
| Business/Enterprise | Efficiency-oriented. Highlights speed, automation, and specific use cases. |
| Call to Action | Minimalist. Uses simple imperatives like "Get started" or "Download." |

## Lexicon
*   **Use:** Introducing, Advancing, Built to benefit, Research, Safety, Workflow, Ecosystem, Provenance, Alignment, Deployment.
*   **Avoid:** Revolutionary, Game-changing, Magic, Incredible, Best (unless referring to a specific metric).

## Messaging do's and don'ts
*   **Do:** Lead with the action or the entity (e.g., "An OpenAI model has disproved...").
*   **Do:** Categorize information clearly using labels like Product, Research, or Safety.
*   **Do:** Use specific nouns to describe users (e.g., "every role, tool, and workflow").
*   **Don't:** Use exclamation points or emotive language.
*   **Don't:** Personify the AI beyond its functional capabilities.
*   **Don't:** Use "we" excessively; focus on the research or the product as the subject.

## Evidence
*   "Built to benefit everyone: our plan" (Mission-driven, authoritative)
*   "Advancing content provenance for a safer, more transparent AI ecosystem" (Technical, safety-focused)
*   "Codex for every role, tool, and workflow" (Functional, inclusive)
*   "An OpenAI model has disproved a central conjecture in discrete geometry" (Academic, precise)
*   "Get started with ChatGPT" (Simple, direct CTA)