The other week, someone told me the best part of being in between jobs — besides putting the fun in funemployment — is having zero work emails. As someone who is constantly playing email catch-up, the sentiment hit a nerve. When I brought it up with friends, they agreed, telling their own tales of email woe and the gloriously sunny freedom that comes with leaving your current place of employment (and work inbox) behind, if only for a week or two before starting again elsewhere.
The black hole that is a work email inbox can feel like a mind-sucking vortex with a gravitational pull that threatens your very existence. Responding to an email, then responding to the response, and responding to the response to your response, can take up so much time it begs the question: Should we all quit our day jobs and become professional emailers?
It's something there should be a solution for, and tech companies, the ones who made email widespread in the first place, are trying: This week, Google targeted the never-ending anxiety of email at its developers conference. On the big stage at Mountain View’s Shoreline Amphitheatre, the company presented a new feature rolling out to Gmail users later this month, called Smart Compose. Smart Compose is an AI-powered tool that works kind of like predictive text does on iPhones. Except instead of suggesting a single word when you start typing, it suggests phrases and even sentences.
Smart Compose doesn’t go so far as to write the entire email for you (dare to dream), but when you see it in action, you’ll see that it comes pretty close, suggesting common phrases and sayings as you type. Before you stand to cheer at the speed and ease with which you’ll be able to plow through your inbox, however, it’s worth taking a moment to pause and consider: As much as we love to hate email, is there a downside to this upcoming communication convenience?
Smart Compose’s arrival in your inbox might seem sudden, but it isn’t unexpected.
“It’s such an obvious extension of what’s been happening for many years,” Karen North, a professor of digital social media at USC’s Annenberg School, said when I asked if she was surprised by this week’s announcement.
The short history of predictive text — that is, tech companies figuring out what you mean to say before you do — dates back to the beginning of autocorrect in the ‘90s. Microsoft filed the first patent, paving the way for plenty of ducking mistakes to come. Other milestones have followed: Google created its search engine’s autocomplete in 2004. In 2014, Apple hailed the release of its “smartest keyboard ever” when it unveiled iOS 8 with text suggestions: “iOS 8 knows your text messaging style. It knows how you email. It knows who you’re writing to. And it knows what the conversation is about.” (Yes, I agree that sounds a little terrifying.)
Google preceded Smart Compose with its own predictive text tool, Smart Reply, which is currently available in Gmail. Smart Reply suggests three quick responses to an email and, as its name suggests, becomes smarter and more attuned to your style the more you use it.
Smart Compose is a little different. The first version rolling out to Gmail users is not personalised based on the sender’s writing style, although that is being considered for future versions, Paul Lambert, a product manager for Gmail, told Refinery29.
Instead, “the Smart Compose model has been trained on billions of common phrases and sentences, not your previous emails,” Lambert said. “That said, when you compose a new email or reply to an existing email, it will take into account the subject line and offer contextual sentences if it makes sense.”
“The whole world of autocorrect [and] smart reply is an amazing one that combines machine learning from your behaviour with a phenomenal use of data that these companies have about the way we all communicate,” North said.
This isn’t a necessarily a bad thing if that data is being used to make communication easier and faster, but there are questions about the depersonalisation that might result. Who really wrote that email that just popped up in your inbox: The sender or AI? Depending on the context of the message, it could create issues.
Take two people in a relationship. You probably don’t want your significant other crafting a message with phrases and sentences suggested to them. The same goes for an email from a potential employee.
"As much as we love to hate email, is there a downside to this upcoming communication convenience?"
North isn’t as worried about depersonalisation or the risk of error: “[Smart Compose] is the quick and dirty ability to reply in real time." But, as evidenced by the massive collection of autocorrect fails, computers aren’t always right, and failing to proofread suggested text could spell trouble. You could make simple spelling mistakes or, worse, contextual ones. At best, you type there instead of their. At worst you might accidentally insert a phrase that does not fit the tone of an important email.
“Back in the ‘90s, people thought of email as a shorthand ability to communicate with each other — abbreviations or typos didn’t matter — and if you wanted real communication you had to send a letter or make a phone call,” North says. “Now, text messages have taken that [shorthand] place, and email is a more formal, letter style of communication.”
Granted, short emails are still sent all the time. But depending on how good Smart Compose is, and how much people use it, it could steer us away from sending longer emails. If anything, you'll need to spend the time you once took to compose a message on proofreading it. (And then have you really saved any time at all?)
In the near future, people will probably just use the technology for the things Google demonstrated onstage: Filling in addresses and quick salutations. (Of course, you don’t need to go with the suggestions Compose gives and can always turn it off if you don’t like it.) But it isn’t hard to imagine a day when the technology, or one like it, will be used for far more.
As North joked in an email to me, “I'm not sure why you didn't just ask Smart Compose to answer this question!”