Note: This piece is part of the Superheroes section of our Opportunity Sprint. Check out the full do-it-yourself AI strategy workshop program here.

AI is learning to see. Until recently, the information contained inside images and video has not been available to computers. But that is changing. Image data is increasingly available to machines which means that there are entirely new ways of using computers to help humans in the real world – extending reality and giving us new eyes.

Humans are highly visual: we devote approximately 30% of our brains to visual processing. We can extract information from images in ways that machines cannot. For example, it’s a trivial problem, even for young children, to describe the content of a photo, summarize a video or recognize a face when they’ve only seen it once before.

As AI gets better at understanding the content of an image, machines can do more; more reliably, at scale, instantly. But challenges remain. AI still requires vast amounts of data, cannot explain why it sees what it does and makes mistakes in ways that are unpredictable or make them too easy to fool, for example, a small change in lighting.

As machines gain more capability in vision, new applications for AI appear.

  • using machines to inspect equipment in dangerous environments or for mundane tasks
  • taking the friction out of check in or check out
  • building 3D models and using augmented reality
  • enhancing radiologists’ ability to spot disease in medical images or enabling patients to better spot disease themselves, say, in the case of melanoma or other skin lesions
  • making cars safer
  • helping companies see when and how their products are used

We think the most inspirational applications today of this machine capability include:

  • using visual search to understand taste and context
  • giving people the opportunity to interact with virtual objects as though they are real
  • making store check out a thing of the past

Watch these videos, discuss what you think is inspirational and imagine what could happen next.

Share on email
Share on facebook
Share on linkedin
Share on twitter