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 can do more than analyze and make predictions—it can go further. AI can discover associations, patterns and even ideas that humans cannot. As AI becomes increasingly specialized at a task, it can be more certain of what it knows, which enables humans to use AI to seek knowledge and truth.

Humans build knowledge by testing theories and trusting others’ advice or opinion. AI builds its knowledge by using its vast scale and by being able to discover correlations that are undiscoverable to humans but that may, in fact, point to something useful. While correlation does not imply causation, AI can be instrumental in enabling humans to understand new causes, new effects and new concepts, especially in fields where there is a full-on data deluge such as astronomy and physics. AI doesn’t think like people; an AI’s knowledge is inscrutable, unexplainable and non-intuitive, which is why it’s so valuable. It’s entirely possible that a considerable proportion of humanity’s new knowledge will be created by AI from data alone, with a human-machine hybrid mechanism for developing any associated theory of causation: observation, simulation and AI-hypothesis generation working together in a virtuous cycle of knowledge creation.

Humans are fallible. In the modern internet, extreme ideas or controversial ideologies are easily propagated. As AI gets better at creating and delivering content for us to consume, it gets better at creating false content or finding ways to trick humans. In some ways, now that AI is able to create images, video, text and voice that can not be detected as “fake” by humans, we need AI to detect AI.

Some knowledge is only accessible via machines. It would take 30 million people to process the information contained in the daily deluge of satellite imagery about the surface of the earth. Clearly we need machines to create the knowledge and make it available to us.

We think the most inspiring examples of AI that develop knowledge and point to the truth:

  • builds a “brain” of who knows what inside a company and helps people find other people that can help them solve business problems.

A short video by Starmind, showing how AI creates knowledge about people in companies
  • can scan the entire internet looking for fake content and tell you how sure it is about its knowledge

It’s an arms race to detect AI-generated fake content
  • can find and manipulate molecules and identify useful new compounds that humans could never find

Microsoft uses AI to make CRISPR even better
  • can search for anything on the earth’s surface

A short video about Descartes Labs, a company that uses AI to help people search the earth

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