According to Jeff Dean, Senior Fellow at Google, no more than a few thousand companies today have the right talent for building Artificial Intelligence, but many more have the necessary data. Based on the feedback I receive in my presentations across all continents, I am persuaded the urgent goal today is to take A.I. from thousands of organizations solving deep learning problems to millions, as soon as possible.
The impact of Artificial Intelligence cannot be underestimated. Millions of jobs will be radically transformed across hundreds of thousands companies around the world. Tasks that still require a heavy human component will be automated, leaving former employees with few options: upgrade your skills or go unemployed. For those who decide to acquire new skills, the sudden increase in productivity will bring a new wave of opportunities that will create jobs we can’t even imagine as of today.
Today firms can deploy Deep Learning frameworks like Caffe, Theano or TensorFlow and not require an army of PhD-educated employees to undertake these efforts. These different frameworks are doing essentially the same thing: allowing firms to write less code to accomplish meaningful tasks. Some of these frameworks are more general than others, but as long as employees can specify the architecture in a framework of their choice, they will be able to get a working model.
In addition to the required expertise in Artificial Intelligence, companies will need to consolidate data currently available as well as conceive innovative ways to start collecting new data. The strength of their models will be supported by the availability of data representative of the real world. That is the most relevant first step in their journey to seamlessly (and profitably) embed Artificial Intelligence into their business operations.