Monday, January 02, 2017

Divided We Fall

Like many others I spent a good chunk of the holidays reading various analyses and post mortems on our national politics and the recent Presidential election.  I am not going to chime in with my two cents on this, which is unusual for me, but my feelings are still too raw and I'm trying to avoid inflicting pain on myself.  Instead I will be taking on what is a safer topic, in the sense that I have little emotional stake in it, but it is still a worry, no doubt.  But I did have a reason to mention those analyses, such as this one by Jonathan Rauch, How American Politics Went Insane.  Some general points from them that I will try to apply in this piece are: (1) we don't know a good thing when we have it; we only can tell in retrospect, (2) in the process of trying to make things better we may very well make things worse, and (3) whatever system we have people will game it and expose weak points that may not be obvious to the developers and the reformers ahead of time.

I want to talk about voice activated virtual assistants.  My wife is now a big fan of Alexa.  She is definitely not the only one.  She learned about it from family friends.  Here is a piece that compares features of the leading competitors, the others being Apple's Siri, Microsoft's Cortana, and Google Assistant.  The piece implicitly is being a cheerleader for use of the technology.  It doesn't say anything about the unanticipated and quite possibly pernicious consequences that might result from use of the technology.

In the process of composing the previous paragraph I received an email from LinkedIn where the featured piece is entitled Just How Dangerous Is Alexa?  The timing of this is almost too much for me.  Clearly, others are worried about the same issue.  And some of those others know a lot more about the technology and security matters than I do.  So I will leave that to them and here do some wild speculation, imagery really, about what might go wrong if other parties have voice recordings of us.  I will borrow from TV and the movies in providing examples.

The show 24 was compelling to watch and I was a big fan.  Part of the attraction is how it was able to inject a dystopian vision of technology and make it an integral part of the story line.  In Day 1, there was an episode where surveillance cameras seem to track Jack Bauer wherever he went, so his remote tormentor, Gaines, could order Jack to do as commanded and succumb to blackmail, with Gaines knowing immediately if his orders had been obeyed.  This type of intensive monitoring is very frightening while at the same time makes for captivating viewing by the 24 audience.

In Day 2 there was an episode which is more relevant to the topic at hand.  In episode 23, a hacker named Max had built this incredible machine to manipulate voices.   All it needed as input was a short sample of the speaker's voice.  Then, as output, it could produce new sentences spoken by that same voice, in such a reliable way that the intelligence services would be convinced that the voice was authentic and not a fake.  If a capability of this sort is even remotely possible, imagine the havoc that it might create should your own voice recordings fall into the wrong hands.

Let's turn to a different movie, The Matrix, and particularly this image, which I think serves as a good metaphor for our relationship with technology companies.



This is meant for us to ask, do the technology companies have our own interests at heart?  A rule of thumb from when I used to do voice recordings at work is that 1 minute of voice takes up about one MB of disk.  That's about 8 year's old information.  Compression is surely better now but maybe the voice recordings are richer as well to offset that.  Does the Cloud storage of voice recordings that the technology company's employ have the right security for it?  Or do the technology companies not get cost effectiveness when they add security layers?  More importantly for this post, how would any of us who reside in one of these pods know the answers to those questions?

The last movie I will use here is from long ago, 1974, the Francis Ford Coppola film called The Conversation, which appeared in the wake of Watergate.  With ordinary technology now (SnagIt) I can record any call I'm part of when done on my computer.  I actually do this with some frequency so that others who are interested in the call but can't make it at the time can view/listen later.  How hard would it be for third parties to make recordings of calls on our smartphones?  I'm guessing, not very.  How would we know if this is happening or not?   We have Joseph Heller to thank for this line from Catch-22: "Just because you're paranoid doesn't mean they aren't after you."  Yes, indeed.

It seems to me we who live in virtual pods have our interests aligned on these issues.  What to do about it, I'm not sure, so maybe we can't yet agree on how to manage things.   But I hope we can agree now that simply waiting for such audio content to get hacked is not the right way to play our cards.  

No comments: