That's not proof in any way that this might succeed now but there is something to be said or giving that a shot with the number of eyeballs available today (and the harder to advertise to markets that have been unlocked by the web but where ad agencies are not yet capable of selling inventory).
If you could centralize that through the browser it would be a fairly powerful strategy if adoption would be large enough (that's the crux), which is why I wondered if it was on their roadmap.
[1], in 1998, I seem to remember that one of my co-workers had installed a thing on his work computer that would move the mouse and "browse" the internet when he wasn't there, getting paid for seeing ads.
Anybody that gets into ads these days will have to take abuse as their #1 priority to be able to compete so if this is to work at all you can bet that anti-fraud measures would make up a very large component of the package.
Tying the end-user and one of the primary beneficiaries to the same chair is a risky thing but not all forms of advertising are equally susceptible to this kind of risk. For instance, performance based advertising and to a lesser extent branding are not all that much at risk.
On another note fraud is present in all forms of online advertising at the moment (with the same qualifications listed above) and it is simply priced into the ad rates.
I don't really believe this project will work, but your argument is rarely a good reason not to try something.
