Any creative solution is built upon at least one new or commonly overlooked or obscure feature of the problem. Datasets for machine learning generally contain the commonly noticed features of things. Access to obscure features is limited. The collection of a thing's obscure features has been proven to be beyond the ability of any computer or human to list out, deduce, or explore in its entirety. So, both computer and human creativity have a limitation.
Better together: Given that humans and computers have different creativity blind spots, they should work together to uncover as many obscure features as possible. Human blind spots have been well-studied and algorithms have been devised to help counteract them. Humans can help counteract the newly proven computer limitations. Together, humans and computers can achieve a state of super-creativity that neither partner can reach alone but they need an interface to record the creative steps of each partner.
Bottom line: Computers cannot completely take over creativity and innovation so many human jobs will remain intact but will be altered.
Other voices in the conversation:
- Jesse Engel, artificial intelligence researcher, Google Brain: Augmenting human creativity
- Simon DeDeo, complexity theorist and cognitive scientist, Carnegie Mellon University and the Santa Fe Institute: thy commitment, decorated with Joy, begins to speak briskly
- Ed Newton-Rex, founder and CEO, Jukedeck: Computers are already creative
- Oded Ben-Tal, composer and researcher, Kingston University: Our definition of creativity will change
- Simon Colton, artificial intelligence researcher, University of London: Machines will be creative for, with and despite us