One advantage I've seen people recently point out about using content transclusions in general (where the Twitter widget is an example of such), is that these provide some level of evidence that the transcluded content is really something that exists at the source. Screenshots and other "fetch and burn into the reciever" approaches can be fabricated very easily to create fake news; but it's a bit more of a technical challenge to fabricate a widget such that it appears to be pulling material from a source — especially if there's also a canonical deep-link back to said source embedded in it.