Artificial Intelligence & Design
Lecture and Workshop by Jae Kim and Austin S. Lee from Microsoft
If we think about design and AI, we could come out with three different ways to connect two worlds. Designing AI, Design for AI and Design with AI. Before we dive into these three different relationship, let’s brief explain about what is AI?
What is AI?
AI has an overarching meaning of computer could do something which we think only human could do. In the computer science field, it is narrow donw to machine learning, program can learn and peforman better overtime. Nowadays, many great accomplishments come from one branch of machine learning, the deep learning which mimic the strutue of neural to construct.
The general approach to use machine learning to solve a problem is that you get a large amount of data and a well-defined question. Then you use the data to train the machine and let the machine answer the question. For AI experts, they need to ask where does the data come from and what question should they ask. If you don’t have the data, you can’t train the machine. If you ask the wrong question, you get wrong answer.
Current AI application actually could fall into one of the categories below.
- Machine Learning — Deep learning
- Natural Language Processing (NLP) — Machine translation, Question answering, Text generation
- Expert Systems — Analytics, Knowledge base, inference engines, Decision making Vision — Image recognition, Machine vision
- Speech — Speech to text, Text to speech
- Planning — Dynamic programming, Reinforcement learning, Intelligent agents
- Big Data — Volume, Velocity, Variety
Designing for AI
Designing for AI means designing for AI service or product. There are many AI products in the market. Siri, Alexa and Google assistant are one of the example. In Microsoft, the animation of Cortana is based on her imaginary personality.
There are so many AI applications and both Microsoft and Google have come out their principles of designing for AI.
Design with AI
Design with AI has two approaches, AI as a tool or AI as a collaborator. We need to ask what do humans do better and where mahince always beat the humans?
Human intelligence and machine intelligence is an incredible symbiosis. — Rob Miller
From
- Everyone will be a designer
- Designers as curators, not creators
- The ongoing era of superstar designers
- From traditional to virtual designers
- Designing AI, designing the future of humanity
Become
- Everyone will have powerful design tools
- Designers as curators, storytellers and multidisciplinary creators of design tools.
- The ongoing era of superstar designers with increase of productivity via AI
- From traditional to virtual designers (Please don’t differentiate the role of the two).
- Designing AI, designing the future of human empowerment.
AI Design Workshop
After the lecture, we have a AI design workshop focusing on the project each team works on.
- Imagining your problem could be solved by current AI technology.
- Imagining your problem could be solved by AI in the next 5 years.
- Imagining your problem could be solved by unlimited technology.
Due to the time limitation, we did not go through the third one. But only by imaging the future technology gave us some interesting ideas.
For current context, our idea is a smart clip which can record conversation between resident and patient. After that student could replay the conversation and practice in the specific scenario.
For 5 years context, we imagined that machine could detect and record even more information including body language and facial expression. While we thought of something like glass, Autsin pointed out that why not consider an non-intrusive way such like multiple sensors in the room.
We only had very short time to think about context wihtou technical limitation and did not come out with anything yet. But it seems to be extreme hard to ideate withou limitation but still make sense.
