Using Gmail auto-complete emails on your friends is just wrong, sorry – Science News
The internet has given us endless new ways to communicate, and endless new ways to make each other feel bad while doing it.
Entering the fray are Google’s optional Smart Compose and Smart Reply features. As well as predictively completing sentences, they offer Gmail users three pre-written responses to choose from based on the email they received.
Send me a photo of your vacation and I might reply “Love it!”, “Looks beautiful!” or “Wow!”.
At first glance, these tools and others like them seem to recognise that answering emails is mostly busy work of debatable value. Google says it “saves you time”. The effort of getting back to people can be a burden — one that the company’s advances in machine learning can alleviate.
But what if you receive what’s clearly a rote response to a note that you dedicated time and emotion to composing? That might not feel good at all.
Do we owe each other more when we communicate? Is Smart Reply not, in its own words, “Awesome!” but morally awkward?
I asked a philosopher, a relationships expert and a technologist to prioritise communication over convenience and give this an ethical once-over.
What is technology there for? According to Matt Beard, a philosopher at The Ethics Centre, that’s the question we should start with.
Often the focus of technology is to increase human capacity — to make what we do more efficient and effective.
Smart Compose clearly makes quick work of replying to emails. But it’s often assumed that if something is more efficient, then it’s more effective. That cannot always be said for communication.
Another field that offers some answers is known as affordance theory.
While Smart Reply may aim to encourage efficiency, it also has little clues or “affordances” — like offering three possible email replies that only take a click of the mouse to use — that prompt us to engage with it in a particular way.
“We would call that kind of affordance an ‘encouragement’, as we’re discouraged from writing an email in the old-fashioned way because it’s more complex [than clicking Google’s response],” he said.
While digital technologies have allowed us to discover communities online, to find new friends and romantic partners, Dr Beard suggested the user experience design of many digital platforms nevertheless make it easy, even too easy, to dismiss people.
Take Tinder’s “swipe left” feature for unwanted suitors, and now, perhaps, clickable email answers.
In a time when online miscommunication is rife, maybe we need more friction, more thorniness to take responsibility for our words, and a little less efficiency.
The relationship expert
When you’re in a relationship, the ways you communicate can start to shape your identity and the nature of the relationship itself.
Are you a couple where emoji are as good as text? Are you more likely to send a picture than words to someone you care about?
Jenny Douglas, a couple and family therapist with Relationships Australia, said knowing you’re receiving computer-generated email responses from a friend could start to change the dynamic.
Too many times and you might feel dismissed. You might assume the other person is too busy to write full sentences. Perhaps for both reasons, you might stop inviting them over.
“For the most part, we don’t like automated responses,” Ms Douglas said. “Outside the convenience of them, most people don’t feel respected.”
It also depends, of course, on context. If you’re making a dentist appointment that’s one thing, but if you’ve been having an intimate conversation with someone, suddenly starting to use obvious canned responses could feel outright insulting.
As the sender, there is also the possibility you stop processing information.
“The inherent risk is you’re not actually reading the communication,” she explained. “If you’re just giving a response that is almost automatic it doesn’t have that intentionality.”
On the other hand, Smart Response and Smart Reply could act as enablers for people who have trouble writing.
“One would argue that any response is better than nothing,” Ms Douglas said.
“But where the technology is enabling, as opposed to a substitution for more intimate or direct modes of communication — that’s where you’re more likely to have difficulties arise.”
In Dorian Peter’s family, Gmail’s machine learning tools are already in use.
While her partner loves it — “it speeds up his routine work emails that only ever required impersonal single-phrase responses anyway,” she said — she was surprised to see it recommended for more personal interactions.
The creative leader at Sydney University’s Wellbeing Technology Lab, Ms Peters said she might feel offence if someone close to her had clearly used the pre-written response — that, as the receiver, she was not important enough for a thought-out message.
“I use Gmail for personal communication and when I first saw Google suggesting I should auto-respond to my mother I was pretty taken aback (as a daughter and as a mother!),” she wrote in an email.
There’s also the unpredictable impact of substituting another person’s or a machine’s language for your own.
Ms Peters described the “personality” of the Smart Reply responses as very “American Silicon Valley”, with their use of words like “Awesome!” and abundance of exclamation points. (Google says the tool should improve the more you use it.)
Settling for its words and personality “smacks of a strange cultural imperialism perpetrated over time”.
Lastly, choosing one of three possible answers is not, she argued, the same as translating authentic thoughts and feelings into our own words.
To that, all I can say is: Awesome, thanks! Thanks! Sounds good.