In the US, go with "visually impaired." That's an accurate description without any baggage, and it also tells you what needs to be done to make the technology accessible to them.

It also describes people who might have limited sight now, but will have full sight later (recovering from surgery, etc). And that is something that could happen to any of us, so we should be building all our tech to cover at least common edge cases, if resources don't permit covering all of them.