AI Isn’t Taking Your Job. It’s Taking Pieces of It.

The boring parts go first

A year ago, I thought most of the AI talk was noise. Then came February 2024. That's when Klarna said its AI assistant was doing the work of 700 customer service agents: two-thirds of their support chats, handled by software, with satisfaction scores unchanged. That tells you exactly how this works. The robots aren't coming for your whole job. They're coming for the boring parts of it. And the boring parts, added together, are somebody's salary.

Then everyone showed up at once

It started in November 2022, when ChatGPT opened its doors to everyone. Within roughly two months, about 100 million people were using it every month. Nothing like it had ever grown so fast. Twelve months on, regular people had it writing their emails, planning their holidays, doing homework, even drafting angry letters to landlords. The excitement faded. The habit didn't.

Text on a screen goes first

The jobs going first all look alike. Same kind of work every time: text on a screen, same steps, over and over. Support chats. Data entry. Routine translation. Product descriptions. GitHub Copilot came out in 2021 and now sits inside code editors, finishing programmers' sentences for them. The programmer still has a job. But the dull half of the day? Gone.

Hollywood saw it coming

Then Hollywood walked out. Writers were on strike for 148 days in 2023, and AI was one of the main things they were fighting about. Nobody there seriously believed a chatbot would write great television. Because studios wanted AI first drafts and cheaper humans to clean them up. The writers saw the real business model before most office workers did: keep the person, shrink the paycheck.

It's already in your pocket

And it's not just work. Just look at your phone. Half the time it knows what you're going to type before you do. Your photos app finds your dog without you labeling anything. Your bank texts you within seconds when someone buys something in another city. Nobody announced that the AI era had begun. It just showed up as software getting smarter, one update at a time.

Google made it official in May 2024 with AI Overviews, those generated answers sitting on top of search results. Fewer clicks to actual websites. For bloggers and small shops that lived off Google visitors, that stung.

What AI still can't do

So what survives all this? The stuff that's hard to automate because it's human, not technical. Talking down an angry customer. Deciding with only half the facts. Being the one who answers when things go wrong. A nurse with a frightened patient. A manager choosing who to hire. A plumber with his head under your sink. You can't automate being accountable.

My advice is boring

Here's my suggestion, and it's not exciting: pick one AI tool and use it on your real work for a month. You'll learn fast which parts of your job it does in seconds. Those parts are on a timer. Where it fails, where it talks nonsense with total confidence, that's your territory. Spend your energy there. Get sharp at the things it can't do.

The people who'll struggle most aren't the ones competing with AI. They're the ones pretending it isn't there.

So, What do you think about it?