Project 2

Description

For this project, you will produce a digital design work that critically engages with the project's theme. You will also develop a design concept, production process, and a well-articulated work statement to explains your work to a wide audience. This is an opportunity to work on a project over a longer period and focus your creative interests in anticipation of the capstone course (DMD-400).

Project Topic: A.I.

Algorithms have shaped our relationship to culture ever since social media feeds and streaming recommendations became dominant venues for creative work. But so-called artificial intelligence tools like image and text generators are evolving at exponential rates and offer new possibilities to artists and designers, while simultaneously threatening entire creative fields with the possibility of automation. So let’s think about what this means for our future, and see what benefits current AI tools can bring to our workflows.

If you're not super familiar with generative AI tools, explore the links and examples further down this page before completing the reading and videos below.

Required Reading/Watching

You will be asked to respond to these in an upcoming blog post.

1️⃣ Creative Dialogue series by Runway (2023–2024)

  • Part 1: Claire Evans & Stephan Marche (6min)
  • Part 2: Anna Ridler & Lex Fefegha (6min)
  • Part 3: Golan Levin & Claire Hentschker (6min)

These optimistic videos were produced by the AI video company Runway, and include some cool examples of generative image and video production. Each video features a great conversation between two artists who have worked with AI and other forms of generative technology, offering some inspiring ideas on the creative process and the role of automation.

2️⃣ Why A.I. isn't going to make art by Ted Chiang. (New Yorker, 2024)

Alternative PDF access(right-click and save as)

This is a more skeptical take on AI art from a renowned sci-fi author. Chiang compares it to historical technologies like photography and drills into the ideas of intention, averages, and authorship.

Project Instructions

Use something generated by AI in your project.

This could take the form of text, images, videos, music, or other media assets that you insert into a creative project.

Or you could us AI tools as part of your process: to brainstorm ideas, discover visual inspiration, layouts, narratives, etc. For example, you could generate some AI poster designs, and then create higher quality posters yourself using colors and layout ideas from the AI-generated images. You could generate a list of possible design solutions to a UX problem, or a short story to illustrate with original comic artwork.

Beyond process or materials, what should your project be about? That’s up to you, but you could consider the topics of artificial intelligence, automation, labor, robots, consciousness, human-ness, or other ideas related to the process we will use. Often when the concept of your work aligns with the form, sparks happen and everything becomes stronger.

These topics are not required; you can choose any subject for your project and simply use AI tools to generate something.

You should still be the primary author of the work and spend appropriate time on the project every week. If AI generates everything, then you will need to justify what value you added as the human director. (This is a pressing question for many industries right now!)

Resources

🗓️ As of August 2024 (New tools or updates may be available)

Any text, images, or media you put into these tools may be used to train and improve AI programs. For that reason do not share overly private information or media; if you have concerns, contact your instructor.

Text

Premium Choice: ChatGPT-4: A paid upgrade on the free model, and you can use plugins to let it browse the web, execute code, and interact with third party APIs.

Free Recommendations:

  • ChatGPT: Fast and easy to use, can generate code; cannot browse the web to find latest information.
  • Claude: Competitive with GPT-4 and notable for having an insanely large context window, meaning you can paste in entire books for it to transform, summarize, or interact with. Maybe a bit more creative and thoughtful than ChatGPT, can generate but not run code, and a beautiful interface!
  • Bard: Google's AI chatbot can run code and "see" images to analyze them.
  • Google Docs: You can generate and edit text directly in Google Docs. To request access, sign up for Google Workspace Labs at the link above.
  • Lex: An AI-assisted text editor, if you prefer that to a chat interface. Write a document and let the AI continue your ideas, generate text, or develop notes into full text.

Example propmts:

  • "How might a mobile app help library visitors engage with books in a new way?"
  • "Write the copy for a landing page about an eco-conscious clothing company. Include section headings and use fun, pithy language that would appeal to teeangers."
  • "I'm developing a fantasy video game, list some potential names for a magical island inhabited by ghosts, using elements from ancient languages."
  • "Who are some artists working with animation and site-specific projection?"
  • "Help me write a comic about a conversation between a robot and a butterfly. Use script formatting, and include a visual description of each frame."

Remember that you can follow-up your prompts with more specifics to "steer" the response in the direction you want. You might have to help the chatbot along a few times to produce the kind of results you're looking for.

In the discovery phase, you can also tell a Chatbot to ask you questions to provoke introspection, as Geoffrey Litt demonstrates in his blog post, ChatGPT as Muse, not Oracle.

Images

Premium Choice: Midjourney: A fantastic tool that is worth a month of subscription to try it out. Uses a simple interface via Discord, where you type commands to generate or modify images. Follow their quick start guide and read about the many commands and parameters available to customize your results.

Free Recommendations:

  • Bing Image Creator: Uses DALL-E model, and is pretty great and easy to use.
  • Adobe Firefly: Notable for the fact that its AI model was trained exclusively on images Adobe owns (through Adobe Stock), so it is not exploiting artists by copying their style, if that's a concern for you.
  • GetImg.Ai: A variety of image creation tools, including the 2024 Flux model.
  • Visual Electric: A Figma-like interface to let you generate lots of images and arrange them together, remix them, and define styles. Free account gets 40 images per day.
  • Playground: A powerful GUI for generating images with Stable Diffussion; 1000 free images and licensed for commercial use.

Example Prompts

  • "A photograph of a man lounging under a tree on a summer afternoon, golden hour, cinematic bokeh, sepia colors."
  • "A minimalist cartoon illustration of watermelon hugging a tomato, vibrant pop-art colors, radiating lines of energy."

Use very descriptive language, references to specific art styles, camera terms, or art media. You can alter the prompt if the results aren't to your liking, and remember that Midjourney has very specific parameters to help guide the resulting images.

Most of these tools will generate 4 images at once, and you can pick your favorite one to scale up to a higher resolution.

👉 For more examples of prompts and how generated images can be used in visual design, check out this blog post by DMD faculty Benjamin Andrew.

Other Image Tools:

  • DALL-E 3: Owned by OpenAi, who makes Chat-GPT, this tool offers a limited number of free images before you have to pay for them.
  • Artbreeder: Combine multiple input images and other visual remix tools.
  • Vizcom: Generate realistic imagery through a combination of sketching (!) and text prompts. Popular with industrial and product design. Looks 3D but I think it just outputs images.
  • Hugging Face: A vast library of user-created tools built on popular machine-learning systems, including some visual interfaces for the open source Stable Diffusion.
  • Spline Style Transfer: Spline is a cool web-based 3D modeling app, and while they're still working on a true "text-to-3D" model, you can use text prompts to generate images based on your 3D models, trying out new color or aesthetic ideas in an instant.
  • Icons8 Upscaler: Most generative image tools create fairly small images, so this AI tool can be used to scale them up them to larger sizes.

Advanced Use of Stable Diffusion:

As an open-source model, you can do an awful lot with Stable Diffusion if you're comfortable with some coding, including integrating it with other applications.

Video & Sound

Premium Choice: Runway: A host of tools, including many for transforming video clips: remove background, enhance audio, remove objects, add subtitles, text-to-color-grade, text-to-3D, bokeh simulation, and more. Now with "text-to-video" generation and the ability to add brief motion to still images! (Limited use with a free account.)

More:

  • EBSyth: Generate video and animations based on an input video + an input image. See FAQ page for tutorials.
  • Mubert: Generate instrumental music tracks online or using a plugin for Premiere and After Effects.
  • Pollinations: A collection of experimental AI tools with a free trial and then paid results; includes text-to-music and video generation, tools for animating still images, and more.
  • ElevenLabs: Generate realistic speech form text, including the ability to upload recordings to use as a basis for the synthetic voice.

Pro Software

Deliverables

  1. Production blog (Canvas)
  2. Concept pitch (Canvas, production blog)
  3. Production Reports (Production blog)
  4. Work statement that uses design language (Canvas)
  5. Project files: (Canvas)
  6. Project documentation (Canvas, production blog)

Schedule

  • Week 7
    • Read Project 2 description
    • Schedule one-on-one discussion with instructor
  • Week 8
    • Concept pitch slides are due (Canvas).
  • Week 9
    • Discuss pitch with instructor
    • Production Report #1 due
  • Week 10
    • Writing Project due
  • Week 11
    • Production Report #2 due
    • Mid-production Instructor Check-in
  • Week 12
    • Production Report #3 due
    • Work statement draft #1 due
  • Week 13
    • Production Report #4 due
    • Work Statement Draft #2 due
  • Week 14
    • Production Report #5 due
  • Week 15
    • Final Submissions:
      1. Project Files
      2. Final Work Statement
      3. Project Documentation

Project media

If you are unsure of what types of digital projects are available to create, please have a look at the list of digital design project types. This is not a comprehensive list, but you should be able to locate your own creative interests within.

Rubric

All project deliverables must be met for the project to be graded. A non-submission or non-completion for any part of the project is considered an incomplete project.

  • Met deliverable requirements: 20%
  • Concept: 20%
  • Research and Process: 20%
  • Craft: 20%
  • Writing and Presentation: 20%

results matching ""

    No results matching ""