Here are my thought on the topic...
- Publishing seems to be moving towards (bandwidth-intensive) video. Blogs, etc... may move towards video, and text and image recognition might make the contents of a video searchable. For example, in the future, if you search for "i love you", the videos containing that sentence will show up (along with a reference to the correct position in the video where it is heard) without the uploading having to actually tag the video himself, thanks to speech and image recognition. Youtube is already experimenting with the techlology.
http://www.labnol.org/internet/video/youtube-speech-recognition-search-text-in-videos/3885/
- Along with voice recognition, hopefully AI will finally pick up and we will have smarter, more human like interfaces thanks to advances in computing and language research.