Prototyping with AI 🛠️ 🤖
Get inspired by some of the references and tools we use. 🧠 ⛈️ 🤓
Imagination of Things was invited to give a workshop with the AI Business Consultant students at Hyper Island, providing them tools and tricks to consider machine learning in the core of the ideation process. Specifically, we sought to remove the misunderstandings, mysticism and magical thinking when incorporating AI into a product idea. What are the questions that can help unlock our perception of AI when creating, developing, implementing business ideas? What are the questions we should be asking?
We started by provoking the students to think about that just that. They wrote down their first questions, things like: “Can we make AI ethical in an unethical society”, or “Can AI help us to learn about yourself more?” From there, we took them through several steps in ideation and prototyping, and here we will share some of our references and tools.
References 📖 💻
Examples of business imagination with ML
Fitness & Body
AI-Driven Fitness:Making Gyms Obsolete?
VAY AI
Objectifier
Art
Perception Engines
Neural Zoo by Sofia Crespo
Infinite Skulls by Robbie Barrat
Allison Parrish
Sam Levigne
Community
Feminist Chatbot by COMUZI
DoNotPay
Music
Magenta Studio
NSynth Super
AI-generated Travis Scott Song
How AI helps Spotify
Roast your own Spotify
Product & Design
AI designs Balenciaga
One-of-a-Kind Labels for Nutella Campaign
Mackmyra: The World’s First AI-Created Whisky
My first generation ML Model with Runway
AI Logos made with Emblemmatic MarkMaker
Climate
PlantSnap
SilviaTerra
Formulating a Concept📝
Participants were prompted to grab a post-it and reference link to create a card that fulfills these prompts:
What is it (ie system, app, law)?
Give it a name and a goal.
Define the role of the machine and the data being used.
Prototype Roadmap 🗺️
We put together simple templates on Miro for participants to use as a launchpad when getting those initial ideas down. We always find that ideas are best understood when there is a step by step of how it is used and experienced and how it works. For this we always encourage images with people in action. Found images, with a couple interventions (drawing, icons, etc.), help to show the inner-workings behind the experience.
Rapid Prototyping with Machine Learning 🤖
We went through several no-code (or minimal code) online tools and demos that can help anyone understand how ML will actually incorporate into their idea, removing the mysticism and black box effect. Coming up with a prototype plan helps to rationalize the use of ML in the invention.
Toolbox 🛠️
Neural Networks
AI Duet
TensorFlow.js
VisualML
performance_rnn
Runway
Lingua Franca
Natural Language Processing
Natural Language Google Cloud
InferKit
DialogFlow
AutoML Natural Language Beginner’s Guide
Hugging Face
Computer Vision
TensorFlow.js Real-time human pose
PoseNet
Teachable Machine
Snap Lens Studio
Spark AR Studio
Image Creation
GAN Paint
AutoDraw
sketch_rnn
Quick, Draw
Image-to-Image Demo
Neural Style Transfer
Nvidia Gaugan
Runway Model Training
Hope this comes in handy! 👋
