Our technological advancements, though astounding, still fall short. There are some algorithms which work wonders, and they will give results for quite a while in the future. This includes things like Google's search technology and context sensitive ads. But most others, which deals with lots of content, based on relevance and is referred to by millions of people do fail. Such failures are immediately brought to attention of the public, and then starts the whole campaign of 'is the source reliable?' which we have become so used to by now. A recent instance of this happening is the news about the 15 year old Tom Vendetta being hired by Google (news article). As it turns out, it was just a juvenile prank which ended up revealing a gaping hole in the system. It showed how blind [sic] algorithms can be.
Websites like digg help to tackle some of these problems, giving human beings the power to control the flow of news and deciding what is relevant. Techcrunch once surveyed that sites like Digg actually control the internet traffic and the pages which people visit. That is probably the best way to go, and is perfectly everything Web 2.0 (which seems to be doing its own guiding these days). Its more sites like these that we need to make relevance more relevant.
The solution as I see it
The best thing to do at this point (until the automations become dependable) is involve some human intervention with the working. Probably keep a team which looks into it from time to time so that the content flows smoothely. If someone can invent algorithms to filter out useful data, they can also come up with algorithms to filter out unuseful data. Based on a rating system, content which falls below a certain rating can be reviewed by someone, and if its found to be actually irrelevant, removed from the system.
Or maybe Google can acquire digg (hehe!) and integrate it with their own Google News, so that news content is posted by both people and their computers, to be rated by other people. Basically, the more we get people involved, the better it will be for us. Atleast, at the present moment!