Data Science and Machine Learning

Machine Learning is acknowlegded as a part of Data Science, but will it be able to replace a Data Scientist?
There have been several articles around that topic in the last few years and months. It’s true there has been some major progress in the field of machine learning and there are already articles about the beginning of automated science like Lipson and Schmidt.
During the SXSWeek there will even be a Panel concerning this topic The Data Scientist Will Be Replaced By Tools.
The main question is, will machine results replace human expertise? There are several startups that provide data science as a service, like Prior Knowledge or Platfora. These companies help to discover knowledge hidden in a company’s data. PK looks in the provided data for correlations and helps to build predictive models. Pltfora on the other hand wants to make Hadoop usable for everyone.
These companies can help discover information, but only combined with human expertise from inside the company it is possible to make the most of the uncovered information. So, in my opinion, machine learning helps making the job of a data scientist easier, because he can concentrate more on his expertise with the context the data was created in.
This may even help in broadening the access to data science to more people.