I'm not saying you're wrong, but "does this technology have potential for abuse" is not a great metric IMO. E.g. computers have potential for abuse, phones have potential for abuse, pretty much any tech we've ever developed that helped someone also coincidentally helped people we wish it wouldn't, whether it be criminals, or just made the government more powerful (if that's something that worries you).

I’m not face-blind, but I do suck at remembering who people are. I want it for my Google Glass so it tells me who I’m talking to so I stop embarrassing myself when someone greets me - now if only wearing GG in public was socially acceptable...

To be fair, that could easily be implemented in a personal manner with manual tagging. There's no need for giant centralized databases or violations of privacy to achieve your usecase.

I was referring to facial-recognition in general - I could train a recognizer on my phone's contacts photos. I didn't mean getting a GG app that used a database trained on public Facebook and LinkedIn photos.

What about something like: extremely dangerous, wanted criminal is on the loose. He's known to be a in a large city, and you use facial recognition to find/track him before he can commit a bombing or something like that.

He said that "didn't have potential for abuse", I think the situation you gave defines abuse. Tracking everyone to catch one person is not acceptable in my opinion.

Yeah but what about: Some guy you really hate for his speeches hides in the city. Why not go there and suicide him before he goes and tells the people how their lifes can be much better or nonsense like that.

Or related to that: Let's assume you really hate the Jews for, I dont know, existing maybe. Why not track them all down and put them in camps? It's not like they can hide from you you, you know where everyone is all the time after all.

That’s exactly the situation that has the potential for massive abuse. There is a central institution that’s tracking all movements of all people. What’s stopping them from going after inconvenient people or selling information?

Visual recognition of individuals seems useful as authentication in a lot of scenarios, in conjunction with secrets. I’d feel better about my bank or title company authenticating me visually for certain operations in addition to current methods. Though maybe a simple trusted video recording without recognition would do just as well for resolving disputes.

Complete blind assumption: Gov LEO tech Probably way ahead of us. The status of a problem doesn’t change need to address and resolve.

Casinos were some of the earliest users.