Researchers still have to define meaningful questions, validate data, inspect methods and decide whether an apparently successful result is actually useful.
The likely future is not science without scientists. It is science where researchers spend more time supervising increasingly capable computational collaborators.