The Mashable blog makes a good point commenting on user-generated content (UGC). 99% of UGC is nowhere near the quality expected of professionally created content. However they recognise that by applying ‘next-generation AI and content processing algorithms’ the good bits can be found. This makes perfect sense. Repeat customer always needs an acceptable risk/reward-ratio for media services.
The user risks time, money or both. Harnessing the power of masses these risks can be minimized. Listening to popular music and watching award-winning movies minimizes the time part. However you could argue that sometimes art and even experiencing it is more about the journey than the end result. Shifting through volumes of content can be fun, especially if there is a real chance of finding something really good. The promise of social media is that by sharing this information an intelligent agent can help you and others minimize both risks.
Same concepts apply to automatically captured video also. Machine-generated content (MGC) services just increases the number of content items substantially. The workflow needs to be very well tuned. Efficient preprocessing and short media duration will help, but even then a very active community with simple and effective tools is required. Challenge is enormous, but that just makes finding the solutions interesting.