Some background and a disclaimer. First the disclaimer, I've worked at Google (which lives of its search advertising) and now I'm with Blekko (which generates revenue from advertisements as well) so I've been on the 'inside' looking out at this problem.
The background. A lot of the web you see today is automated, which is to say that 'the web' can exert influence in the 'real' world, and influencing the web can be amplified (or leveraged) as traders would say, with automation.
Two examples, in the first example we have a thief who wants to steal money but doesn't want to go out and rob a bank. Instead he 'hires' Google (an unwilling partner here) and creates an AdSense account, makes a web site, fills it with AdSense add javascript, then writes a program that can fetch pages from his web site and click on an ad. He then feeds that to a botnet for hire (generally not an EC2 cloud) and has it go off and click on all his ads. Google congratulates him on what a great web site he launched, pays him anywhere $10,000 to $50,000 from ad revenue. And the party continues until the Advertisers complain or the traffic patterns pop out as suspicious. Obviously Google has a lot of folks who do nothing but look for this sort of abuse but when you're making numbers like they, even if you had 7 or 8 million dollars a month in click fraud it would be rounding error in their quarterly reports. The people who get pounded by this are the Advertisers which lost their $500 or $50 or whatever that month by being charged for clicks from machines that weren't ever going to buy their product.
In the second example we have an ambitious entrepreneur who wants to get the word out on their new thing. They create over time a few hundred thousand 'fake' twitter accounts, they construct a community of fake accounts following each other. Then they write a simple program to have their 'influential' tweeters push out the word and their 'followers' pick it up and retweet it. If they have let the accounts lay there for a bit they pick up their own share of robot followers (from other people doing this) and our ambitious entrepreneur creates what looks like a real groundswell of interest in what ever thing they are trying to push.
There are many rewards to automating web activities, for many different reasons. As a service provider its a full time job trying to insure I can identify the fake ones (I don't want to send ads to robots, they don't buy anything and they make me look bad to my advertising channels) That 80% of the 'clicks' this person got on Facebook fail a simple 'real user' test does not surprise me.
I expect there to be a rise of landing pages which are themselves pages requiring you to click somewhere else and then clients only paying out when a prospect clicks all the way through to some part of the web site. I know that some big names already do this where they bring up a page asking you to identify your location (which helps them bring you to the part of the site for your area) but they also don't pay for the click unless you make it through that page.