Dylan Moses responds to a recent paper arguing that ML outputs lack 'speech certainty' - being identified with certainty by the speaker at the moment it's spoken - and that such outputs thus lack First Amendment protections. Moses proposes a simpler solution to the question about ML and the First Amendment: "ML output shouldn’t receive First Amendment protections at the moment of generation because it is not human. The Supreme Court has never protected the speech of non-humans absent some connection to a user or receiver." Moses' reply to Levy & Austin is available from SSRN.