Last year Gmail introduced a feature called “Smart Compose”, which uses artificial intelligence to predict what you are trying to say. It’s in many ways similar to the Apple predictive key keyboard, which has appeared on a while back, which also gives you suggestions of words that it thinks are appropriate in the context.

Basically, algorithms are getting better at predicting what we are trying to say. It makes writing that slightly formal email easier – one is able to resort to the stock phrases of politeness without actually having to think too much about it.

I may be a little bit late to the party here, as I typically have predictive keys switched off on my phone. As a bilingual speaker, I tend to find that most of the time my phone completely fails to recognize what language I’m typing in. Hence, a feature most people seem to find indispensable, I tend to find at its best useless, at its worst downright bloody annoying. I’m sure someone in Silicon Valley is working on this problem of automatic language recognition right now (or may have already solved it).

But this is just the tip of the iceberg of my reservations towards predictive texting. As someone who likes to think of herself as a writer, I would like to think of my writing as original. Predictive texting seems to brutally expose the fact that more than half our writing is routine and can be produced by an (admittedly rather complex) algorithm and some statistics.

We like to think of ourselves as unique in some way, but predictive writing seems to prove that we tend to say pretty much the same things.

Instead of emailing each other perfectly original sentences, like “The dog wagged its tail and swallowed the universe in its yawn”, we spend our time thinking we are being perfectly unique but not being that unique at all “What’s the project deadline?” and “The meeting is canceled.”( If you’ve somehow missed this, here’s a fantastic Fry and Laurie sketch about language https://www.youtube.com/watch?v=3MWpHQQ-wQg , which inspired this reflection).

Basically, I can just about tolerate predictive writing in a work email. But when a computer is able to predict my creative writing, I will probably plunge into a full-blown existential crisis.