Friday, September 21, 2012

Bad Argument: abusing crowd-sourced filtering

One of the powers of the Internet is the ability to poll the crowd to know what's good and what's not. This is known as crowd-sourced filtering. The best known version is Web of Trust, i.e. WOT. You can leave ratings and comments on any URL you go to.

The hope is that people will use it responsibly, put real honest ratings, and thus, the "accurate" opinions will outweigh the inaccurate ones. However, this may not happen immediately, and sometimes, when ratings of a suspect scheme drops below a threshold for it to be consider "scam", a bunch of supporters will then flood the ratings with positive reviews to hopefully push it back into "neutral", or even "good".

Here's an example..., one of the biggest Ponzi scheme to his the US in recent years, has a low WOT rating.

However, look at their comment statistics:

Yep, almost 400 people left positive comments, whereas only 14 people left negative comments. All of the positive comments say the same thing: great way to make money. In fact, this attempt to influence the vote was caught by WOT admins:   Seems most of the comments hit WOT right about August 2011, when Zeek changed the comp plan to require "giving away bids" to earn that RPP.

Similar things happened at Norton Safe Web. All of Norton Safe Web's comments (4 of them) was added 3 months ago (June 2012)

This tactic is known as astroturfing, where a suspect scheme tells its members that its reputation needs "defending", so they go make positive comments about the "opportunity". The fact that they BENEFIT from such reputation defending, of course, is not mentioned.

The negative version of astroturfing is known as black propaganda, where a smear campaign is used against opponent by filing a flood of fake reports and low ratings, false report of malware on site, even false claims of copyright violations, takedown notices, and so on. Skeptics' websites are sometimes targets for such tactics when they raise issues with pseudo-science promoters. My hubpages hub about ZeekRewards was hit by such a bogus takedown request that knocked it offline for six days.

[ read more about Internet reputation management ]

What can you do about such a thing? Other than encourage "good" people to vote and monitor the rating systems for such abuses not very much. But it is still a "dirty trick" you need to be aware of. "Good" ratings are not necessarily good, and "bad" ratings are not necessarily bad.

The point is "trust, but verify", skeptic's motto. Even if the info is real, you need to know HOW did the info came to be, how it was collected, and so on. If someone's "massaging" the data, you want to know about it.

Enhanced by Zemanta

No comments:

Post a Comment