AI Usage & Disclosure

01 Our position on AI

AI is a tool. Like a press, a font foundry, or a stock photo library before it: powerful, useful, with limits. It accelerates ideation, draft work, and the boring parts of production. It does not replace creative judgment, strategy, taste, or the ability to look at a piece of work and know it's right.

Every project we deliver involves human direction at the start, human judgment in the middle, and human review at the end. AI may participate; humans decide.

02 Where we use AI today

Drafting and ideation

• ChatGPT (OpenAI): brainstorming, copy drafts, summarization, light research.
• Claude (Anthropic): longer-form writing, document analysis, strategy work, internal operations.

Image and design

• Canva Magic Studio: in-app image generation, copy assist, photo edits within Canva templates.
• Adobe AI tools: Firefly for image generation, Generative Fill in Photoshop, Generative Recolor in Illustrator, and related Adobe Creative Cloud AI features.

Code and build

• Lovable: AI-assisted web development. Currently powers parts of cmykink.io itself.

Operations

• Otter.ai: meeting transcription and summarization. Used with participant consent on every recorded call.

03 Where AI never decides

• Strategic direction: what to make, why, for whom.
• Final creative review on any deliverable.
• Client communication. You're talking to a person.
• Pricing decisions and contract review.
• Legal, financial, or compliance judgments.

04 How we use AI responsibly

• We use enterprise or business-tier accounts wherever available.
• We don't paste sensitive client information into general AI tools without explicit client permission and the right vendor agreement.
• We review AI outputs before they reach you. Drafts get edited. Generated images get refined or replaced.
• We disclose AI-assisted work when you ask, and proactively in case studies and educational content.

05 Your right to opt out

If you don't want AI tools used in your project, say so before we start. We'll honor it. There may be efficiency or pricing implications. We'll talk through them with you.

06 What AI can't guarantee

• Outputs can be wrong. We catch most in review. Final approval responsibility rests with you, the client.
• Outputs are not guaranteed unique. We layer strategy, creative direction, refinement, and brand-specific systems on top. That's where the uniqueness comes from.

07 Ownership of AI-assisted work

Copyright law around AI-generated material is unsettled. Our working position:

• Final deliverables go through enough human creative direction and editing that they qualify as our work product, transferable to you per the MSA.
• Pure AI outputs without meaningful human shaping may not be copyrightable in the United States. We don't deliver projects that depend on pure outputs.
• Our internal AI prompts, workflows, and systems remain ours.

08 Third-party AI platforms

AI tools we use are operated by their vendors, governed by their terms and privacy policies. If a vendor changes its terms in a way that affects your work, we'll tell you and adjust.

09 Ethical use

We don't knowingly use AI to impersonate a real person deceptively, create unlawful content, infringe intellectual property, or generate exploitative, harmful, or misleading material.