UGC Video Update - How It Will Be Safe For Advertisers

From the latest OnlineSpin:
"(...) There are three primary ways in which I've heard of companies developing  tools to ensure the safety of advertisers in this space. The first methodology  is based on peer review. In this model, a set of meta-data for the  content--primarily video content--can be created based on the reviews or any set  of review information posted and maintained by the viewer. It relies on the  engagement of the audience, but has proven successful for other sites,  especially those that are in the form of the Wikipedia. This model requires  ongoing maintenance by the audience, which can be viewed as a pro or a con--but  is certainly useful, because as the environment changes the meta-data can be  updated to reflect audience tolerance.  

The second model I've come across is when a technology catalogues the content  on the page surrounding the UGC and hypothesizes that the content of the video  is related to the content surrounding it. This model would seem to be relatively  easy to build, and its logic is clear, but the problem is that it needs to  cross-reference the data for the multiple of pages where the content is located  and then merge the data to ensure proper classification. This process is  extremely important when cataloguing for search. This model would appear to be  sound, even if it requires that a fairly robust database be maintained to ensure  its effectiveness.  

The third model I've heard about is similar to the second, but rather than  the content surrounding the video being catalogued, it's the pages linking to  the video. This seems to be most applicable to YouTube and the other sites where  other viewers consistently link to the video content from other sites, but do  not host the videos themselves. This model also appears to be the most  scalable--although once again it requires that the spiders be in place on a  consistent basis to ensure that the content is updated regularly.  (...)"