Siri Goes “Bonk!”

This entry was posted in Smart Design, Your Pal Drew and tagged , on by .

2013_starcops_001

For the last week or so I’ve been trading messages with my friend JD about the new movie “Gravity”, as we both share an interest in a sci-fi sub-genre referred to as “hard” science fiction (fictional stories which adhere to real-world technology and physics while avoiding the temptation of using inexplicably “magical” technologies). I’ve also been upgrading to a new smartphone all week and a few days ago I had a bit of a technological epiphany… but first a quick discussion of the smartphone.

After more than 4 years of using an iPhone 3Gs I upgraded to an iPhone 5s, and the first thing I wanted to try out was the famous speaking “digital personal assistant” named Siri. My main desire was the continued hands-free control of my music library, but I was naturally curious about Siri’s other capabilities. Over the next few days I plumbed the limits of Siri, and I learned a few things in the process.

The first thing I learned was disappointment.

Siri is in NO WAY what I would consider to be the artificial intelligence that she’s been pitched as in commercials; she’s simply a set of canned answers wrapped around a voice translation routine, with a limited number of hooks into the available software applications running on my smartphone. Any intelligence that people may sense from “her” is due to the natural tendency of humans to anthropomorphize inanimate objects.

 

Siri goes “Bonk!”

Siri doesn’t work without a connection to the Internet, and if you stray from the fairly small list of things she’s able to discuss then you’ll hear one of several cute responses. After a short while those cute responses begin to be as annoying as the loud error BONK! you might hear from a regular computer.

 

Siri doesn’t listen well.

Her voice translation routine doesn’t seem to be as powerful as Dragon Software’s products, and she’s far too quick to decide that any hesitations during dictation indicate that I’m done talking and ready to send my message, forcing me to dictate the message all over again. Hey Apple: pauses don’t mean that I’m finished talking, I’m just southern!! Seriously, there ought to be a control panel for setting the delay period before the software finishes translating a person’s speech to text, or a keyword for indicating that you’re finished dictating. My dad suggested that I whistle during pauses to keep the software from cutting me off, and that has actually seemed to work even if does make me look a little crazy.

 

Siri can’t do much, and some of the things are parlor tricks.

The speech commands haven’t been granted enough power to reach deeper into the phone’s operating system. For instance, I wanted to know how many songs I had in my iTunes library but Siri had no idea what I was talking about when I asked for that fact, nor did she know how much free memory I had left.

All I heard were excuses… “Bonk! Bonk! Bonk!”

I was entirely unsuccessful at convincing her to inform me about dozens of my phone’s settings, but hey – I could have her turn on the flashlight! Woohoo!!

After more than a few failed starts at these types of searches Siri patently struck upon the concept that I must be looking for something online and started offering me links on the web to peruse. The only problem with her offer was that I was driving 55 miles an hour down a highway and she wanted me to click her screen and read the articles myself, even though I’d initiated the search using spoken commands.

 

Siri can’t read.

Every so often she would offer to read me an entry on a topic from Wikipedia, but that’s the only site she’s willing to read from. I haven’t looked this up but my first suspicion was that Wikipedia may be one of the few sites that adhere to strict Web Content Accessibility Guidelines – perhaps one side-effect of an increasing use of spoken assistants like Siri might be a surge in well-structured websites. Or maybe the code team behind Siri are just snobby. Or perhaps there’s a lurking legal issue of repurposing one site’s content for audio transmission by the Apple product. Whatever the reason is, I want Siri to be able to read to me from any site I point her toward.

 

Siri can say dirty words.

Siri can repeat VERY explicit curse terms. The problem is: she doesn’t know that I’m directing them AT HER.

 

Siri doesn’t take insults.

Siri cannot differentiate between being cursed at and being asked for help, so she’ll often begin looking up those bad words for you.

 

Siri can’t even tell you about herself.

Siri cannot provide you with a spoken set of help commands for interfacing with her, instead she puts up a list of potential actions on her screen. Why would you respond to a speech command with only an on-screen response? Apple is missing a great opportunity to highlight Siri’s speech interface abilities; they aren’t thinking like Steve Jobs. Boo, Apple. Boo.

 

Star Cops

So, back to the beginning of this post: there I was, looking up showtimes for Gravity on my iPhone when I was suddenly reminded of an old BBC detective show called “Star Cops“; a late-80s hard science fiction show set between Earth and the moon. The main character uses a phablet-sized mobile computer presenting a remarkably prescient rendition of Siri. Watch the first 10 minutes to see “box” in action… or at least watch the first 3 minutes to hear the show’s awesome theme song by Justin Hayward from the Moody Blues.

So, what do I want from Siri?

Far too much, far too soon, apparently. I want what writers back in 1987 were able to envision. I want her to be like Nathan Spring’s box from Star Cops: a *real* artificial intelligence.

2013_starcops_002

She should be able to parse conversational dialogue and reciprocate with rich, dynamic responses filled with data, not a lot of Bonk-Bonk-Bonk canned responses. I don’t want to have to push a button in order to talk to Siri, she should be able to tell when I’m addressing her and be ready to go. She should be able to read to me if I ask her to. She should remind me of my schedule, even nag me intelligently.

Touch screens are fun, but I’m trying to resist gluing one to my eyeballs. Speech interface technology is a viable future. It’s certainly come a long way in my lifetime, but Siri still feels frustratingly like beta software at this point.

Siri is 2 years old and even though I’ve only had her for a week I’m anxiously waiting for her to graduate to a newer, smarter version. Apple, if you drop the ball on this somebody else will pick it up, and you can bet your ass that I’m going to follow them.

Siri is still handy

Don’t get me wrong: I like Siri. Now that I’ve found the technology’s limits I’ll be able to make it work for me – at this point a box of moldy crackers would have been more useful than my 4 year old iPhone. I’ve really enjoyed the ability to send text messages without taking my eyes off the road or my hands off the wheel, and I’m sure that I’ll bridge the gap of the limits of the technology by adapting my usage.

Postscript: Android’s solutions are really no smarter.

I saved this one for last because I knew that all of you Apple-hating Google-ball-sucking fanboys would be rejoicing while reading this Apple-user’s critical review of Siri. Well, you can go find a different article to validate your Apple-rage because whichever company’s technological wiener you lust after is no better than Apple’s Siri… and to top it off: nobody has even HEARD of your technology. What’s it called, Edna? Mabel, maybe? Probably Ebby Mae. Whatever it is, the point of this post was to show how far we still have to go – all of us, regardless of platform.

3 thoughts on “Siri Goes “Bonk!”

  1. Erica

    Actually, Siri used to be a standalone app, waaaaaay back when- 2009 maybe? I had it and deleted it almost immediately because voice command really isn’t my thing.

    These days, Siri refuses to acknowledge my identity (per recent FB post).

    I am fascinated, though, by her pronunciation skills.

    Things Siri has no problem saying correctly:
    Pohorski (a friend of mine)
    Cahuenga (a road in LA, KAHwenga)
    Sepulveda (another la street that confuses A LOT of people, esp. Back-Easters, including me – it’s spelled se-pul-VEY-da, for some reason pronounced seh-PUL-veda … which would be spelled Sepúlveda, but I digress)

    Things Siri effs up royal:
    Los Feliz (yet another street, it’s pronounced in a wacky anglicized way here (las FEEEliz). Siri neither hits that, nor the Spanish pronunciation (los feLEES))
    Hillhurst (wtf? ENGLISH!)
    HORNUNG. No real shock there, also may explain the “I don’t know who you are” type interactions I generally have with her.

    … You drive 55 in the ATL????? I’m calling you OUT, son.

    Reply
    1. Drewprops Post author

      “You drive 55 in the ATL????? I’m calling you OUT, son.”

      Totally busted. I wanted to wrap myself in civic responsibility.

      Reply
  2. John Kirk

    Hi there,

    Very interesting article, yes I do agree that Siri is quite limited but I’m still excited by this early start. I too am a keen fan of star cops, if I could buy a case/cover to transform my smart phone into Box then I would buy it! I’m an author of hard science fiction so check out my website, you may find the ‘inspirations’ section interesting. Enjoy!

    JK

    Reply

Leave a Reply

Your email address will not be published. Required fields are marked *