I do like the thought process though - if it's supposedly this common but still unproven, it's less likely to be true. Do we put ghosts in the same category?
There is still potential for mass-fraud, but repeated seemingly uncoordinated reports improves credibility. Maybe they're all Photoshopping the same thing from different angles, but that is less likely than a single person doing it.
^ I mean come on, if you were an alien, where would you go first? I'd totally go to New York. Here is another rule of thumb: if the aliens are doing something boring, like stealing cows or fucking with corn fields, they probably aren't real.
Hmm. That would be an interesting piece of fraud (oops I mean public art) to architect. Make a Blender scene of Central Park, populate the scene with bystanders with phones (yes, this will take time to source realistic models and skeletal animations). Each phone gets to be a camera in the scene. Not all cameras start and stop at the same time, and a few of them will have to capture video before and after the event without the actual event included. Timestamp and tag the videos appropriately, including clock drift, then have fun with avconv/ffmpeg converting the formats to different video formats based on known phone sales and what default formats iOS/Android encodes to (ugh, I actually had to do this once back when zencoder/pandavideo).
Create twitter accounts in advance, make them look real by copying other twitter users, and the day of the event just start uploading the youtube videos all at the same time and get your fake tweets flowing.
Now, what could this pipeline be useful for? Suppose you know that the videos are fake, but also suppose you can be reasonably sure that others will accept this data as real news. What could you do with that? First idea is some killer marketing stunt..
I think maybe you could do it by actually filming in Central Park with some sort of structure with white balls on it, in a certain known arrangement, that you could automate a lot of the video editing from various angles.. I'm not sure about that but I suspect it is feasible. If you get everyone to film different angles at the same basic time then the scene would be the same (just from different angles) and you'd have everyone else filming in the background that would lend credibility.
Cameras are not very good at catching sharp movement of low lit objects. Just to photograph sports clearly requires several thousand dollars in equipment and you can be as close as 50 meters most of the game.
That will change a lot over the years as sensor technology continues to improve and eventually your point will be important.