Just read the post about AutoML, which is a very good introductory post about the new Automatic Machine Learning field.
Per the discussion in the post about one of motivation for AutoML is that release the hand of data scientist to do more challenging stuff, which I think it is a little bit distort the fact that it is also reducing the employee pool of data scientist.
This is a soften version of saying I want to replace you one day for your own good. I generally do not oppose the AutoML. But I do think the entry level of data science job that the people on the position cannot understand the core idea of machine learning algorithms will be endanger of their jobs.
The future of AI and data science will be the people who super good at implementing or improving by change the core of algorithms. Meanwhile, the communication of the job is also a key that these AutoML pipelines need to be maintained and interpreted by DS.
