I don't think that's true. An entity (corporation, government, etc.) continuously decides whether to humanize or dehumanize its citizens/customers. Which way it leans in any (every) decision is a matter of goals and values, not a matter of scale.

E.g. a government may not know all its citizens in a personal sense, but it can establish charters of rights and freedoms, it can design humane and accessible services, etc. Scale doesn't prevent these things, rather it necessitates them.

Humane and accessible services are still designed at scale, rather than on an individual level.

Of course. In context, I read "anything at scale does this" to mean "anything at scale dehumanizes its users" which is the point that I disagree with. A scaled-up service can still honour the client -- it takes intention and good design, but it's certainly possible.