Really, Gmail?
To mother, with love
The spread of AI has brought with it a creeping realization about the way we communicate at work. At root, AI is a tool for generating vast amounts of Inoffensive Generic Prose (which we’ll call InGeP). The fact that it is sweeping through the world of work as rapidly as it is, and embedding itself into more and more of our daily working lives, strongly suggests that work—and in particular white-collar work—relies more on the generation of inoffensive prose than we might perhaps have realized. All the emails back and forth; all the seven-bullets-with-seven-words PowerPoint slides; all the biz speak and performative “excitement” about this, that, and the other: for better or worse, much of modern business is conducted in a language that prizes the smooth surface and the generic formulation—and speed.
So when AI offers the automation of InGeP, business grabs hold, hard. As a result, a cheery button inviting you to generate a little dollop of it has materialized on most of the applications you use every day. Beyond this, even, the generic prose is now materializing directly, often in places which were previously InGeP-free zones. After my post last week, a friend contacted me to tell me that Gmail now proposes a response for every single incoming message—and, as you’d expect given the current state of AI, does so with distinctly mixed results.
In my friend’s case, it had learned her pet name for her mother but then mashed this into a tone-deaf comment about her mum’s illness. “Bummer about the CT scan, Ma,” said no sentient child, ever. But Gmail was right there.
Which is...what? Weird? Spooky? Icky? Whatever it is, it’s certainly not what we want from our tools. The idea that AI is effectively lurking in the bushes outside our homes, eavesdropping on the private names we call one another, is downright creepy. And the idea that someone thinks you can outsource emotional labor to a machine is just plain repulsive.
(Also, what’s the ideal use case here—where is all this heading? My AI chatting away to your AI? What problem are we solving? That it’s sometimes hard to find the right words? Is that a bug or a feature of a caring human?)
When we add this to the unwanted intrusion of ClippyPilot that we talked about last week, we can see that it’s entirely possible that AI is much less than it’s cracked up to be.
But there’s another possibility, and that’s what I want to raise today: that AI, before very long, will be much more than it is now. At the moment, it is a fountain of InGeP. Because of this, we can rant about it and lament its intrusion, and get irritated at having to listen to the tech bros and their Promethean pronouncements about the future of civilization. And there is an interesting discussion to be had here, too, because at the moment AI is a deeply imperfect tool that is simultaneously fascinating and addictive and may well be quietly manipulating how the world works.
But what if it gets better, as there is every indication it will?
It’s entirely possible that in the blink of an eye, AI will be a fountain of good prose. And if Inoffensive Generic Prose is useful in certain contexts (the transactional email, say), then what will the advent of automatic and actually good prose mean for work, and for humans in general? It’s easy to dismiss a crappy Gmail suggestion (and please, Gmail, spare us). But will a good one be so easily dismissed, and if it becomes good, should we ever dismiss it?
I think the answer is yes, and importantly so. There are more reasons to resist automated word-generation than the quality of the prose—and it behooves us to start talking about what those things are now. Here’s a starter list.
When what matters most is that a communication came from a human, we should never outsource it to AI. This is what is ultimately most grating about my friend’s story above: even a perfectly-crafted AI response can never be a substitute for a message written by a daughter for her mother, because not-written-by-daughter and written-by-daughter are fundamentally different things.
Where the friction of creation is an essential part of the process of making something, we should never outsource it to AI. The removal of friction is not always a good thing, and it’s certainly not a good thing when the friction itself leads to the insight. As tempting as it is to use AI to remove friction from our lives, we need to be judicious about this.
Where what we seek is something fresh, something distinctive, something that stands apart from the pack, we should never outsource it to AI. Uniqueness and distinction do not follow from averaging all the other ideas in a particular space.
And where it matters most to us that we did something with our own hands, we should never outsource it to AI. It will too quickly and happily steal our pride of authorship if we invite it to. Even if we reach the point where the mother can’t tell the difference between the email written by her daughter and the email written by AI, the daughter can.
That’s certainly not an exhaustive list. It will probably expand over time, and you might have some different things on yours. But most important is not whether your list matches mine. It’s that each of us—now, as AI and its Inoffensive Generic Prose shows up in more and more places, and tempts us with more and more instances of the easy button—create our own list. What will we never outsource, irrespective of the quality of the automation?
My list tells me that when it comes to the words, it’s not just the words. What goes into our communications is at least as important as what comes out. The effort to find the right words, the friction of creativity, the imagination required for empathy: these are what make us human. Thinking, caring, and relating can never be outsourced. But their trappings can, and if we surrender the shapes of our humanity, then our humanity itself will surely follow.
In addition to writing about work, I advise businesses around the world on leadership, performance, and people. If you’d like to explore how I can help your organization, please check out my website here.


I just saw a meme that both humors and frightens me: "Right now, the dumbest person you know is being told they are brilliant and correct by AI." This is honestly what is irritating me most about this moment. Really smart people are being pushed to the middle and uncreative people are being pushed to that same middle. Humanity is becoming averaged in a very unhealthy way.
Brilliant, again an analysis with insight and realism!