Automating cognitive tasks: fact or fiction?

The arrival of cognitive computing in the mainstream of business conversation is beginning to create disruption in how enterprises engage with potential future strategies and even with current project planning.

Managers charged with business development now look at products like Amazon’s (amazon.com) Alexa and Google’s (google.com) Home and see not just consumer-oriented home entertainment products, but also AI factories churning through billions of data points of internet voice to revolutionize the relationship between families and “computing” (not to mention between advertisers and their “targets”).

The late, irreplaceable management guru Peter Drucker always enjoyed diving into the heart of the matter—root causes. One of his pronouncements was (I paraphrase): “In the Industrial Economy, the primary challenge for the business was to automate the task. In the Information Economy, the primary challenge in business is to understand what the task is.” When information rules, when both products and processes are made up increasingly of information itself, organizations and individual workers must acquire the imagination and the innovation capability to figure out how the business can stay (or become) successful.

A new type of economy

I would suggest that we are in the early phases of yet another kind of economy, an Intelligence Economy, and that this shift will cause us to rethink the primary challenge for businesses once again. While the naming committee has not yet struck the phrase that will describe the central challenge in the new environment, it will certainly include the need to understand and to define the cognitive tasks we want our machine intelligence to carry out for (or with) us. It will also include new imperatives for ethical practices in a world where machines operate increasingly on their own, but that is a subject for another day. My focus here is to take a closer look at what constitutes a cognitive task and to ask whether we are in a good position to manage the process of engaging machines to do that work?

We casually talk about the arrival of machine intelligence, but what kind of intelligence (or intelligences) are we expecting of these machines?