The Design Emporium Policy on AI

I have been thinking about this and talking to other creatives about this since someone made a series of images showing a capybara suing Elon Musk in 2022.

My thoughts on AI change constantly as I use these tools and experience use of them, so I keep this place as an overview on how I feel about them and how we incorporate them in creative work.

Our standard snippet in standard proposals, which guides the specifics:

We believe in tools that enhance creativity, not replace it. We use AI at the Design Emporium to research, test assumptions, refine concepts and automate repetitive tasks but your final product and the strategy it supports is created by human hands specifically for your project and your objectives. We will never sell you a design that was created by artificial intelligence, because we believe it might compromise originality and industry ethics.

We see a distinction between Efficiency AI and Generative AI.

Examples of the ways we use artificial intelligence:

  1. Early comparative landscape research: finding every peer or competitor in an industry niche so we can understand who is out there and who we are speaking to

  2. Filling gaps in knowledge: writing in assumptions and prompting an AI tool to ask questions that I haven’t considered asking so that I can fully understand a new space

  3. Building efficiency-based tools: Modernized spell check supported by AI and not a static dictionary. A widget that checks color accessibility within design files. A widget that creates pairings of every possible brand color. Using an AI tool to organize 200+ exported logo files into their respective folders.

  4. Crafting outlines: Companies know that a blank document is daunting, which is why Google Docs is always dangling Gemini at the beginning of a document. I might use AI to create an outline for something I’m working on to prompt myself, but then write from there.

Examples of the ways we do not use artificial intelligence:

  1. Hands-off methods: “Make me a logo” or “make me a flyer”

  2. Generative Copy: I’m never going to send you an email or a report written with AI. Proposals are scoped and written by hand (but I do check my pricing against peers based on industry standards, see #2 above).

  3. Providing feedback and direction: I’ve seen clients use AI tools to provide feedback because they feel like they can’t put their opinions into words. Just send us what you would write into Claude and we can understand.

  4. Summarizing Emails and Transcripts: If you wrote it or said it, I want to hear it from the source. Otherwise, what’s the point of writing an email at all?