I would not generalize from Watson to the current trends in AI.

I had researched into Watson when it had won Jeopardy, as much as I could publicly find, and concluded at that time itself that it won't stand up to the hype IBM is creating.