By Michael Andor Brodeur GLOBE CORRESPONDENT MAY 23, 2016
Whenever I meet one of those couples who habitually complete each other’s sentences, I think two things: “Wow, these two are made for each other.” and “Wow, these two need to split up as quickly as possible.”
It’s an irksome quirk — advertising intimacy while telegraphing suffocation. It feels like watching two people mug each other, back and forth. You’re torn between stepping in and stepping away. Worst of all, it’s an insufficiently tacit confirmation that despite the limitless expanse of human thought, our paths across it are ultimately rather predicatable. It’s cute, I suppose; it just bums me out.
But since I’m psychoanalyzing, I might as well consider the possibility that perhaps this tendency resonates so negatively with me because it too closely resembles a corrosive relationship of my own: Me and my phone.
It started back in 2014 (i.e. iOS 8), when Apple first introduced QuickType, an amped-up version of predictive text that lets you forgo typing by tapping suggested words it pulls from a custom dictionary, culled and crafted from your past chats. Thus, the words it suggests are attempts to capture your unique style and voice, your tics and lols. It felt intrusively helpful. (Is phone-splaining a thing yet?)
But as many have discovered for themselves, if you experiment a little and restrict yourself to tapping only QuickType suggestions, the result is a hauntingly abstract self-portrait — a hybrid of confessional and generative poetry. Here’s one I just tapped out in a text that royally confused my husband:
“I’m at a time when the first place I love the fact is the most recent version and other options and a lot to be in my room for a week.”
Like a little puddle of gibberish, it doesn’t run deep – but on its surface, you can barely make out a distorted reflection. It might not make any sense, but it speaks to the power and potential of this fast-advancing technology.
And if Google’s new “smart messaging” app catches on, our phones will be even more keen to speak for us. Allo, which debuted last week at Google’s annual I/O developer conference and will be available this summer, features a number of amped-up standard chat-app functions: You can send stickers and animations, you can doodle on photos, you can shrink or enlarge your text with a swipe, and you can easily summon Google’s new digital assistant to carry out various tasks right from within a chat — that’s the “smart” part.
But the other component of the AI-enhanced Allo is Smart Reply, which, according to a blog postby Google, will allow you to “respond to messages without typing a single word” by learning from your conversational patterns. Allo can discern “whether you’re more of a ‘haha’ vs. ‘lol’ kind of person,” but it can also manufacture text responses to images: “If your friend sends you a photo of tacos, for example, you may see Smart Reply suggestions like “yummy” or “I love tacos.”
Yummy? I love tacos? (I do, but still. . .) Immediately, I’m freaked out and defensive: Get out of my head, Allo!
Rudimentary forms of predictive text have been with us for decades. Those first formative spellcheck sabotages perpetrated in 1997 by Microsoft Word not only inspired a catch-all term for assistant-assisted errors (the “Cupertino Effect,” coined when Word would ironically zap a hyphenless cooperation into Cupertino, the home of Apple), but a whole genre of comedy (see:Damn You, AutoCorrect). And anyone who lived through the pre-smart cellphone era remembers fondly the tap-saving efficacy of T9 (“text on 9 keys”) shorthand — it was the only thing keeping thousands of RAZRs from getting thrown at a thousand walls.
Early efforts in predicitive text seemed like genuinely offered solutions to practical problems posed by the technology itself or the fatness of one’s thumbs. The idea behind Allo, on the other hand, seems to be offering assistance where it’s not necessary — like coming home to find your intern matching up your socks — and it feels a little invasive.
Text messages are among our last sacred spaces in the digital realm. Within them we speak freely, and let tiny textual idiosyncrasies signal our humanity through the screen. Do I really want Google hoovering up everything I say? Do I need it to put words in my mouth? Faced with Allo’s predicitve powers, will I go out of my way to throw it off course? Say “OK” one day and “okey-doke” the next, just to prove to Google (or myself) that I’m more complex than its best guess? This relationship already sounds exhausting.
Like everything else Google comes up with (save for maybe Wave), I’m sure I’ll eventually warm up to the eager pupil that is Allo — maybe once it can convincingly bang out my column? For now, if Google imagines I want my conversations reduced to a menu of prefabricated expressions, it doesn’t know me so well.