We Automated the Boring Stuff. Now What?

Aurora Monroe

There’s a particular kind of anticlimax that’s started showing up in conversations about AI, and it goes something like this. Someone uses a tool to draft an email, summarise a document, or plan a project in about forty seconds. It works. The output is fine, maybe even good. And then they close the laptop and feel, unexpectedly, a little hollow.

Nobody talks about this part much. The productivity wins are easy to quantify and satisfying to share. The hollowness is harder to explain, and it sounds ungrateful besides. You asked for a faster way to do the thing. You got a faster way to do the thing. What exactly is the complaint?

This piece is an attempt to take that feeling seriously, not as a rejection of AI or automation, but as a signal worth paying attention to. Because somewhere in the process of removing friction from our lives, we may have quietly removed something else too.

The Bargain We Made

The promise of automation has always been the same at its core: handle the tedious stuff so humans can focus on what actually matters. It’s a genuinely appealing idea, and it’s been repackaged and resold with every wave of new technology. The dishwasher would free up time for family. The spreadsheet would free up cognitive load for strategy. Email would free up the delays of physical post. Generative AI would free up, well, nearly everything else.

Each of these things delivered, in its own way. Dishwashers did reduce the time spent at the sink. Spreadsheets did remove enormous amounts of manual calculation. And yet at no point did most people end up with the glorious surplus of time and mental space that each innovation seemed to promise. The hours saved got filled. With more tasks, more expectations, more communication, more output. The baseline shifted upward and the breathing room closed.

This isn’t a new observation. Economists and sociologists have been noting the paradox of labour-saving technology for decades. But it’s worth restating as we absorb a genuinely new wave of automation, one that reaches further into cognitive and creative work than anything before it. The bargain is being offered again. It’s worth being a little clearer-eyed about the terms this time.

What Friction Was Actually Doing

Here is the thing about friction that the productivity conversation tends to skip over: some of it was useful. Not useful in the sense of making tasks take longer, but useful in the sense that the struggle itself was doing something for the person doing it.

There is a reasonably well-established body of research on what psychologists sometimes call productive struggle, the idea that working through difficulty, rather than around it, tends to generate deeper learning, stronger memory retention, and a more durable sense of competence. The effort leaves a mark. Not just on the output but on the person producing it.

A related concept is the IKEA effect, named after the furniture company and the curious fact that people tend to value things more highly when they’ve had a hand in making them. Even if the finished product is objectively identical to one assembled by someone else, the one you built yourself carries more weight. The labour invests the object with meaning. Remove the labour and the meaning goes with it.

Think about what this means for some very ordinary tasks. Writing a first draft of something, even a fairly routine piece of professional communication, involves a process of clarifying your own thinking. You work out what you actually want to say by the act of trying to say it. Planning a trip from scratch, including the slightly annoying bits with transport connections and accommodation trade-offs, leaves you with a mental map of the place before you’ve arrived. Filing a tax return is tedious, but by the end of it you have a working knowledge of your own finances that you simply don’t get from handing it over.

None of this means that all friction is worth preserving. Some of it genuinely is just waste, repetitive and joyless and offering nothing back. The distinction matters, and it’s one that tends to get flattened in conversations about efficiency.

The Skill Erosion Question

Somewhere around the mid-2010s, a small but persistent body of research began suggesting that GPS navigation was affecting people’s ability to find their way without it. Not dramatically, not in ways that were immediately obvious, but measurably. Regular use of turn-by-turn navigation appeared to reduce activity in the parts of the brain associated with spatial awareness and route learning. People were getting where they needed to go. They were also, gradually, becoming less capable of doing so without help.

Navigation is a convenient example because it’s relatively contained. The stakes of losing your internal sense of direction are low, mostly. But the same mechansim applies to skills where the stakes are considerably higher.

Consider the situation facing junior writers, analysts, or researchers entering workplaces where AI tools are now embedded in standard workflows. A significant portion of the beginner work, the drafting, the summarising, the initial research pass, is exactly the kind of task that AI handles well and that organisations are increasingly likely to automate. From an efficiency perspective, this makes complete sense. From a skill-development perspective, it presents a quiet problem. That beginner work was never really about the output. It was about the person doing it becoming gradually less of a beginner.

The long-standing model for developing expertise in most fields has involved a pipeline: you start at the bottom, you do the foundational work, and over time, through accumulated repetition and feedback, you build towards something more sophisticated. If AI handles the bottom of that pipeline, it’s not obvious where the foundational skills get built. You can’t skip the beginner stage and arrive at expertise. You can only skip it and arrive at a capability gap, one that might not become visible until something goes wrong.

Purpose, Identity, and the Work We Tell Ourselves We Do

This is where things get a bit more personal, and a bit harder to articulate without sounding melodramatic.

A lot of people’s sense of who they are is tied up in what they can do. Not in a grandiose way, but in the quiet, everyday sense of taking satisfaction in doing something competently. The writer who crafts a sentence they’re proud of. The analyst who spots a pattern no one else noticed. The project manager who holds a complicated situation together through sheer organisational skill. These things are not merely tasks being completed. They’re moments of being good at something, and that feeling, small as it often is, is part of what makes work feel worthwhile.

When a tool does a version of that thing faster and with less effort, it creates a mild but real dissonance. The task still gets done. The output might even be better. But the moment of competence, the part that felt like yours, has been handed off. Over time, across many such handoffs, the cumulative effect can start to feel like a kind of hollowing out. Not a crisis, not a breakdown. More like a slow rearrangement of what work actually consists of, and a nagging uncertainty about where your contribution really begins.

This isn’t unique to AI. It’s a version of something people have felt at every stage of technological change. But the speed and scope of what’s currently shifting means that a lot of people are experiencing this rearrangement simultaneously and without much of a shared vocabulary for describing it.

The Counter-argument Worth Taking Seriously

It would be a mistake to let this piece tip into romanticising difficulty for its own sake. Not all friction is meaningful. A significant portion of the work that automation replaces is genuinely grim: repetitive, intellectually unrewarding, and draining in ways that benefit no one. The data entry, the formatting, the boilerplate, the administrative loops that eat hours without leaving anything behind. Getting rid of that is not a loss.

More importantly, the people who stand to gain most from AI-assisted capability are often those who’ve had the least access to the kind of support that makes hard things easier. A first-generation university student who can’t afford a tutor. A small business owner without a legal team. A non-native English speaker trying to produce professional written communication. For these people, AI tools are not removing meaningful friction. They’re removing barriers. That’s a meaningfully different thing, and any honest account of automation’s effects has to hold that alongside the concerns about skill erosion and identity.

The risk, then, is not automation itself. It’s the uncritical application of it, the default assumption that faster is always better and that removing effort is always the goal.

So, Now What?

The question is not whether to use these tools. Most people reading this already are, in one form or another, and the broader direction of travel is not reversing. The more useful question is what kind of relationship to have with them.

One framing that seems worth considering is the idea of intentional friction. Not as a political stance or a productivity hack, but as an occasional, deliberate choice to do the harder version of something because the doing of it matters to you. Writing a first draft before asking AI to improve it. Planning the first leg of a trip yourself. Working through a problem before reaching for a tool that will solve it in seconds. Not always, not even often, but sometimes, when the skill or the thinking or the feeling of having done it yourself is the actual point.

There’s also something to be said for being more precize about which frictions are worth preserving and which genuinely aren’t. The tax return might be one to automate without guilt. The piece of writing that you want to feel like yours, maybe not. The distinction requires actually thinking about what you’re trying to get out of an activity, rather than defaulting to whichever option is fastest.

The Part That Comes After

The automation wave promised us time. It has, in various ways, delivered time, or at least delivered speed. What it hasn’t delivered, and what no tool ever really can, is a clear answer to what that time is for.

That question belongs to the person. And answering it, slowly and imperfectly, through some amount of struggle and uncertainty, is probably one of the frictions worth keeping.

Aurora Monroe
Latest posts by Aurora Monroe (see all)