The present and future of post production business and technology

Artificial Intelligence and Design

Design is much more than the way things look. It encompasses every aspect of every interaction with the device, object or software package you’re using. Josh Clark published a talk he gave on behalf of his design studio big medium, that is focused on design in the era of the algorithm, or Artificial Intelligence.

It’s a very long and deep article, well worth the read if you design anything at all. I found it very relevant as we are currently designing our most complex and powerful app. Read on for a list of the topics.

The topic list provides a simple way to summarize the content:

  • What’s the next mobile (what’s the next thing that will similarly transform our culture)
  • Technical advance inflates expectations faster than it can fulfill them
  • Machine Learning powers it all (even at Apple)
  • Welcome to the Algocracy ( Algorithmic interfaces are a huge part of our future, and getting their design right is critical—and very, very hard to do.)
  • Machine learning on the cheap (API access is easy and inexpensive, but interestingly the author completely ignores IBM Watson as an Machine Learning API provider)
  • Ready for Experiments (Because stringing together multiple APIs is powerful)
  • The machines make mistakes (The world is a complicated place, full of nuance. And sometimes machines aren’t terribly subtle.)
  • Google and the one true answer (In trying to get the definitive answer, Google can create fake news)
  • Productive Humility for machine-generated interfaces

he then continues with ten design principles for conceiving, designing, and managing data-driven products.

  1. Favor accuracy over speed
  2. Allow for ambiguity
  3. Add human judgment
  4. Advocate sunshine
  5. Embrace multiple systems
  6. Make it easy to contribute (accurate) data
  7. Root out bias and bad assumptions
  8. Give people control over their data
  9. Be loyal to the user
  10. Take responsibility

The article provides a deep look at designing with machine learning, and I recommend you read it.



, ,