Everyone agrees on the goal. Align AI to humanity.
It is the one sentence the labs, the governments, the ethicists, and the public all say together — the rare point of consensus in a field that agrees on nothing else.

It sounds unarguable, the way for the good of all always does.
Which is exactly why it deserves a second look, because the sentences no one questions are the ones doing the most work unseen.

Ask the obvious question and it comes apart in your hands: whose humanity?
There is no single will called humanity to align anything to.
There is a planet of eight billion minds sorted into tribes that distrust one another, nations that arm against one another, faiths and ideologies that define themselves by what they oppose.