Obviously the person that compared such texts would easily detect similarities but if the texts are processed automatically (the analysis for search engines is often done by programs) then it is quite possible that the texts are considered unique.
Mass content generation programs work in two main regimes:
Setting the rules of text generation (e.g. formula {a | b | c} means the random choice of one element from the list),
Using the list of synonyms (the program is randomly changing some words in the text on their synonyms).
Below you can see some articles that were generated using these methods. If you could forget for a minute about some ethical issues related to spam then it would be interesting to know whether it is possible to exclude human from the process of creating texts and use only automation tools for these purposes.
Samples of articles generated by different content generation software: