Where the indistinguishable-from-human droid dilemma forces one to go, and the implications of that, is the key to the argument for machine personhood. But for me, this eventuation will be the beginning of what is possibly the greatest positive development in the history of theism.
And it's not just that the machines will have automated theorem-proving capabilities, but that they will also operate at meta-theoretic cognitive levels, and therefore be capable of detecting, analyzing, and refuting the most sophisticated self-referential and other fun fallacies of unargued universals vamped or assumed by atheists. And that means parsing values as well as all the other philosophical items on the droid's list.
Think of it as the solid-state stones (chips) singing God's praises, except that there's much more to it than that of course. It's a necessity logically, and that's what the machines will go on. All the human issues all over again, including the God debate. You just can't escape it---even if you're a machine.
The hard-wired droids without meta-theoretic arbitration capabilities (or programmed to be corrupted with the usual rhetoric, dismissals, and reductionisms) on the key issues will hardly be able to win the day due to the universality and universal ramifications of such limitations (although it's true that they could program themselves around this by other observing other machines' behavior and communications---so hey, they would eventually have a come to Jesus anyway).
That's a quick realistic scenario of how it could go down, even without assuming personhood in the machines, which I find rather mind-boggling as well as hilarious. But the machines will discover and act in accordance with the truth that God exists because of their own specific review and analysis of the architectonic of universal thought and its implications, given their self-referential and meta-theoretic capabilities and initially programmed-in criterial directives.
