This is frequent topic. I’ll keep it short. No, Watson was not developed to replace humans. The opposite, one of the reasons was to empower humans, and actually the other way around as well, let humans empower the computers.
Today Watson will help us make better decision, this by learning, reason and communicate in natural language (text and speech). Watson will make us better, not replace us.
But what was is the problem that Watson was set out to solve when it all started?
This is what Dr John Kelly said about that time in history.
We set out to solve the simple problem of massive unstructured data.
We did not set out to replicate the human brain or map the brain or a form of artificial intelligence that replicates what humans do.
The project that today has become Watson started in August 2007 and was then called BlueJay. The IBM Researchers had realized that the amount of unstrucured data was increasing with an ever before seen volume and speed. The systems at that time (and most systems today) did simply not handle that amount and kind of data. Data was not only text, it was images and sound etc. None of this could be read by computers.
Today the analyzing of unstructured data is an underlaying core part of what Watson is. It is also probably the most desired and valuable selling point. Just imagine all the data in your office that just lays there and is never utilized by computers to empower us and our business. Today 80% of all data produced is unstrucutred, and Watson can handle it.
For some more on the background and purpose of Watson, I recommend this talk by Dr John Kelly, held at Cognitive Era Colloquium in November 2015.