Sunday, August 7, 2016

Siri not Jarvis

Where's my e-butler?

Advancing computing technology seems to have deviated from the speculative sci-fi books that I read as a child.   I haven't decided whether it's a positive or a negative thing.

I distinctly recall reading stories where households had servers with storage, computational power, etc.  Specifically, I'm thinking of those where middle-class families had an AI butler, or secretary.  They could be relied upon to accept verbal (voice) commands, questions, and possible conduct conversations.  

They'd do more than turn the lights on and off.  They might question whether you were accepting visitors if someone knocked on the door, and either allow them in or politely tell them you were indisposed.  They might accept standing instructions to deny all visitors/calls, but still be smart enough to pass through calls deemed to be emergencies.  The AI would take voice dictation memos (emails), send them off to their destinations, and search various databases (internet) for answers to questions, or research topics.  

It would be a like a butler/servant/personal-assistant, but without the moral baggage of being human. Rich aristocrats may have no problems ordering their "lesser" humans around, but us middle-class folks know that it's morally suspect... since all people should be equals socially.   The e-butler gets around this problem. They are nearly as smart/trainable as a person, but their whole purpose is to serve.  ...and because we made them, we can be sure they don't have their own pesky motivations.

Admittedly, we do have some of these now.  There's Amazon Alexa, Microsoft Cortana, Apple Siri, etc.  One could imagine that they are developing towards "good-enough" AI status.    But there's nothing private about the conversation one has with them.    I don't need to worry as much about a hacker/corporation/government surreptitiously interrogating my butler.   I could trust that they would be savvy enough to know when to guard my privacy, or inform me that they were getting questions.

With Alexa, I can be pretty sure that my inquiries about prenatal development are going to lead to more advertisements for diapers.  ...or that asking about the caribbean will lead to advertisements for vacation packages.  More nefariously, one could imagine that questions relating to politics might prompt a visit from the authorities in a police state

So instead of getting a personal assistant that might have some kind of loyalty to us, we get a communal 'public' assistant who is helpful, but likely has loyalties to someone else.

I suppose this is part of a larger trend towards everything becoming "a service" rather than a thing.

Edit:

Related articles:

  • https://www.theguardian.com/technology/2016/jan/31/viv-artificial-intelligence-wants-to-run-your-life-siri-personal-assistants