In a nutshell, ScreenPlay encourages better software engineering practices, and thus makes the tests faster to write and easier to maintain in the long term, but requires a better understanding of OOP to yield its full benefits (though it does scale remarkably well with junior engineers as long as there are a couple of senior engineers around who understand the pattern).
Page Objects where only ever intended as a stepping stone to help introduce automated testers to better test automation practices, but it can still lead to hard-to-maintain code. Take a look at this article for the origin story: https://ideas.riverglide.com/page-objects-refactored-12ec3541990.
There are a couple of other implementations of Screenplay, including:
- JNarrate: https://bitbucket.org/testingreflections/jnarrate/wiki/Examples
- Narrative: https://github.com/tim-group/narrative
ScreenPlay is an actor or user-centric model, which is easier to scale and more reusable than step libraries in many cases. Once a task is written, you rarely have to change it (respecting the Open-Closed Principle). An actor-centric model also makes it easier to model situations where several users interact in a scenario (for example, with different browsers). Jan Molak talks about the actor-centric side of Screenplay in this article: https://janmolak.com/user-centred-design-how-a-50-year-old-technique-became-the-key-to-scalable-test-automation-66a658a36555
