Can AI Actually Take Over Web UI Development Jobs in the Near Future?

Frontend
1

I keep hearing everywhere that artificial intelligence is about to replace web developers who work on user interfaces. People also say backend programmers might be next. This talk got louder after the new O1 model started doing well in programming contests.

But I tried using AI for real projects and I’m not so sure about this. I asked ChatGPT O1 to build a simple rating carousel where users can swipe through different reviews. The results were pretty bad. The positioning was wrong, the auto-scroll broke when someone clicked manually, and somehow a 4-star review kept showing as 5 stars.

This made me think that AI still can’t handle the tricky parts of frontend work. Things like getting pixels exactly right, making smooth animations, and handling all the ways users might interact with a page. These tasks need a lot of attention to detail.

I noticed AI works better for server-side coding where the logic is more straightforward. But creating good user interfaces seems much harder for AI right now. You need to understand design principles and user behavior, not just write code that runs.

Why do people think frontend jobs will disappear first when AI actually seems weaker at this type of work? Has anyone else noticed this gap between what AI can do and what frontend development actually needs?

2

AI companies demo their tools with backend stuff - coding challenges, API endpoints, that kind of thing. So everyone thinks AI crushes systematic programming tasks. But frontend? That’s all judgment calls. Visual hierarchy, accessibility, UX flows - you can’t just run tests to check if those are right. I’ve used AI for boilerplate and basic component setup, and it works fine there. But throw complex state management or design systems at it? It falls apart fast. People think frontend is easier because they don’t get what modern UI development actually involves beyond slapping together some HTML and CSS.

3

honestly the hype’s way ahead of reality. tried copilot for responsive grid work last week and it kept suggesting outdated flexbox patterns that broke on mobile. ai might get there eventually but frontend’s got so many edge cases and browser quirks that take years to learn

4

that’s really interesting! have you tried being more specific with your prompts about the carousel behavior? maybe break it down - ask it to handle just the swipe detection first, then tackle the rating display separately. i’m curious if it’s a complexity issue or we’re just not communicating clearly with it.