Is working with AI still real work?
I can work on several projects at once now. That’s exactly what makes it harder to call it a day.
Sometimes an AI-assisted test is still running on one project while I’m already planning the next steps for another. When the first result comes in, I switch back and check whether the test actually covered what I needed to know. Then there’s a third project waiting for me to decide what happens next.
Yes, working with AI is real work. In my own projects, AI can get me to a first draft faster. I still have to define the task, keep the systems running, check the results and decide what’s ready to use. Whether AI saves me time depends on that whole process.
Why people have such different experiences with the same AI
If you ask ChatGPT a few broad questions and get three generic answers back, you have good reason to be sceptical. A short prompt is enough for some tasks. For others, I need to explain what I’m trying to achieve, which information matters and what a good result looks like. Sometimes I only realise how vague my instructions were when I read the draft.
Take software development. I can ask AI to change a feature and write tests for it. It’s good to see those tests pass. Then I need to find out what they actually cover. Does the change work with data created years ago? What happens if the process stops halfway through and someone runs it again? AI can help me work through those cases. I decide whether to release the change.
Even with experience, the effort doesn’t pay off every time. If preparing the context and fixing the result takes longer than doing the task myself, the shortcut has cost me time.
When more things can run at once, more things stay open
At work, I can feel how quickly expectations change. When a first prototype comes together quickly, a project can feel further along than it is. Decisions, integration and quality checks still lie ahead. Across several projects, they often land at the same time.
That’s the part I find most tiring. While I’m reviewing one result, I’m already thinking about which task to hand over next. When a process gets stuck, I need to work out whether the technology failed or my instructions were unclear. Sometimes I shut the laptop and my mind keeps going. What’s still open? What should I tackle first tomorrow? Is something already getting away from me?
At times like that, I feel closer to burning out than I’m comfortable with. I need time away when I can genuinely switch off. It’s hardest to take that break when I feel like one more run might finally get something finished. Sometimes I’d be better off finishing one thing properly and putting the other aside. AI makes it very easy to keep both going.
What the sceptics get right
It bothers me when working with AI gets dismissed as pushing a button. I’m equally sceptical of the idea that bringing AI into a task automatically improves the result. A plausible answer can be wrong. A draft produced in minutes can create more work for the person who receives it than it saved the person who produced it. And when you’re unfamiliar with the subject, a confident answer can make a mistake harder to spot.
So I ask whether AI helped with the task at hand, counting preparation, review and rework. I also ask what happened to the time I saved. Did I get more room to think through a difficult decision? Or did I start three more things that I’ll still be thinking about tomorrow?
My conclusion
Some tasks take me substantially less time with AI, and I’m glad they do. For others, the effort shifts. I spend less time producing the first version and more time clarifying the task, checking the result and standing behind it.
How much someone typed tells me very little about how good their work was. I care whether the result does what it needs to do and whether someone takes responsibility for it. If AI gives me an hour back, I’d like to make that choice for myself a little more often. I could start another project. I could also finish work an hour earlier.
Frequently asked questions about working with AI
What work remains for people when AI takes over tasks?
Someone needs to define the goal and requirements, check the result against the actual use case and decide whether it’s ready to use. Running systems also need maintenance and someone to respond when things go wrong.
When does AI actually save time at work?
When the entire task takes less time at a comparable level of quality. That includes preparing the input, reviewing the output and making corrections.
Why can AI increase your workload?
Using the time you save to start more projects can leave you with more open decisions, reviews and handovers. How much extra work that creates depends on your workflow.