Google cannot rely upon dates provided by web servers or content and never has. Let me explain how this works.
When Google finds a page, the creation date (drawing a parallel) where Google calls the date the inception date, is the date and time the page is discovered. If there are two identical pages on the web and Google finds one, that is the original. It has no real way of knowing otherwise. The exception is to use the canonical tag to point to the original work. When a page is modified and Google finds it, the modification date is not the date the actual page was modified, but the date that Google noticed the modification.
These are the dates used and has been the case from day one.
There were experiments to show in the SERPs (search engine result pages) dates provided by the web server or content itself. This may continue in some cases, but was mostly a disaster. However, the dates shown in the SERPs are not the dates stored and used in the index. These remained as they always have been- when Google discovers the page or when Google discovers the page has been modified. This was maintained for continuity. Remember that Google tries to eliminate any way that factors can be manipulated and this has always worked extremely well for Googles purpose.
The idea of freshness does not mean that any page has to be modified. It means that the site enjoys new and fresh content and that any older content be updated as it becomes wise to do. These are two separate concepts. It does not pay to make small insignificant changes to a page thinking it will remain fresh. Google is hip to this trick. Semantic weighting of the page will give this trick away in a nanosecond. Update pages when it makes sense to do so and add content at a reasonable and steady pace. Even one new blog post a week is enough.