The complexity split happened when every game started needing full voice acting to be marketable. Before, text was perfectly sufficient, and a lot more flexible.

When text-to-speech algorithms approach real voice acting, game world depth will start to notch back up.