Unfortunately a speaker's content can also be auto-generated now, at least for brief enough snippets. And that means the content can (and will) be optimized to appeal to a target segment much more than has ever been previously possible.