AOL was it's own network, completely parallel to the internet. It didn't use TCP/IP, it used a propritary transport called P3 - heavily optimized for dialup.

They had their own dedicated client app, where each page loaded in its own window. It didn't use HTML, it used something called "Rainman". URLs weren't a thing, you accessed "channels" (pages) by entering a specific "keyword".

Later on, in 1993, they added support for Usenet (see: "Eternal September"). Then in 1994 they added support for Gopher and WWW. So you could dial into the AOL client, and then open a regular web browser. But for most home users, there was more content in the AOL walled garden, so the web was something of a curiosity at first.

So to answer your question: It wasn't an artificial limitation. AOL was designed in a way that was fundamentally alien to the way the Internet + Web evolved.

(At some point, you could connect to AOL over TCP/IP - useful if you had a broadband Internet connection but still wanted access to content on the AOL network. This was done by encapsulating P3 inside a TCP/IP header. You still had to use the AOL client software and have an AOL membership.)