Identifying Made-up New Sites (aka Fake News)

This is a Public Service Announcement for the upcoming US Presidential Election in 2020. There are a lot of machine-crafted sites out here on the Internet. Learn how to identify some of their features.

Identifying Made-up New Sites (aka Fake News)

Photo by Kayla Velasquez on Unsplash

I recently received a news alert from Google Alerts in my inbox that caught my attention. The site it was from, newswelcome.com, is not well-known to me. I seem to get a lot of alerts from Google for various sites that pretend to be financial analysis sites. For whatever reason, these seem to pop up relatively frequently.

As I looked at the page that Google fed me in my alert, it just struck me as somehow odd. Something wasn't right about it. It read like it was machine-generated, which I am 99% certain that it was. But why? We hear about all sorts of people out there that are either purposely feeding disinformation / lies or sites that are just created as click-bait for ad-revenue.

This article below on a PR firm that was purposely distorting reality to affect an election may be more common than we think:

Disinformation For Hire: How A New Breed Of PR Firms Is Selling Lies Online
One firm promised to “use every tool and take every advantage available in order to change reality according to our client’s wishes.”

When I came across the news site in question, I really couldn't pin-point why it exists. It doesn't seem to have ad revenue on it, it didn't have affiliate links to make some cash off people purchasing items (I do that here occasionally, just to help support the site - $1-2$ a month isn't enough, but this is a small site) and it doesn't seem to be signing up people to sell their information. My only speculation at this point is: could it be used to feed ETF algorithms to influence computer-trading systems?

The page that I hit was:

https://newswelcome.com/2020/01/09/heres-what-to-know-about-verisign-vrsn-stock-is-now-overbought/

The link indicates that a stock may be overbought. Would a computer scraping the site use that as part of its algorithm to make buy or sell decisions?

The page itself reads like it was built from simple variable substitutions. Write some basic text (stock is good, you should buy, you should sell) and then fill it in with some values you can scrape off other financial sites with possibly accurate data. Here is a markup of a few paragraphs from the article:

I highlighted in red some of the values that could easily be dropped into a script to generate all of the text in the article. Grab a stock symbol, pull some data, fill in the values, post. It makes you seem and sound smart, right? Like you did all this research on the stock, so it must be true.

Highlighted in yellow is an assertion the article makes about the stock. It is overbought. This is one element that could easily be used by a computerized trading platform to make micro-decisions on whether to buy, hold or sell stock.

If you read the article, it is readily apparent that the author (again, probably a computer script) is a non-native English speaker. This of course, is not a bad thing, however the photo of the author would possibly lead you to believe they are a native English speaker. The site is all in English. The name used is an English name. But look at the bio:

Right off, you can tell that is not a native English speaker.

Mark Watkins write about Market Movers, as he is a big interest in stock trends.

The image also struck me as odd. Being someone that uses stock images at times (like the lead of this article from unsplash), it felt like a stock image. I ran it through tineye.com as a reverse image search and sure enough, that guy is all over the Internet.

Maybe he does have a big interest in stock trends... along with sunflower-kitchen-curtains, but somehow I doubt it. And that my friends is the real kicker here. Clearly this site is fake. They even went so far as to have an Our Staff page that lists his bio as if he works there (only this time the photo is higher resolution, probably to save on bandwidth because who would look at the staff pages?)

What about the company? They have an address listed with a phone number. The zip code is legit for Ottawa Lake, Michigan, but there is no address 1839 Pride Avenue, as far as I can tell.

So why do these sites exist? I'm not sure I know any more than I did at the beginning of this post. I could speculate all day long about it. Maybe it really is for computer trading algorithms to disrupt the market and make small gains. Maybe it is an English site that is positioned towards non-US readers to influence their stock purchases. Maybe it is a site to make the owners sound financially astute to impress someone else? Maybe it is a way to generate traffic toward the domain name so it could be sold later for a higher price (domains that already have traffic can be valuable for people wanting to jump start a business, users already go there, or in this case Google indexes them and thinks highly enough of them to rank them in a Google Alert email)?

I could go on and on, but the point is, don't believe everything you read and look for tell-tale signs of potentially crafted messaging that might be 100% computer-generated pages.