Introduction

I was recently looking at a fake news article that didn’t have the date the article was published anywhere on the page. I learnt a quick tip a few months ago to remedy this – you view the source code of the page and find the date in there, it normally looks something like:

<meta property=”article:published_time” content=”2018-08-27T21:29:52+00:00″/>

It’s a quick and simple way to find when a web page was published online. When I was doing this for the site clashdaily.com, I noticed this:

document.write(*<div id=”* + c4vl39.mess + *” style=”display: none;background-color: rgba(0, 0, 0, 0.2);color: rgb(255, 255, 255);padding: 10px 20px;text-align: left;line-height: 1.4;font-size: 16px;font-weight: 100;font-family: Open Sans, Arial, sans-serif;letter-spacing: 0.3px;”>\nHany Ad Blockers break the way this site works, causing images to be hidden, videos not to play, content to flow in unexpected ways, and so on. Beyond that, our website relies on internet advertisements to pay the bills, so that you can read it for ‘ );

It’s a piece of code that displays a message when someone is using an ad blocker – this is common across all types of sites. For example, here’s what happens when you view mediabiasfactcheck.com with an ad blocker enabled: In the code above from Clash Daily, one thing that caught my eye was the exact message that would display:

Ad Blockers break the way this site works, causing images to be hidden, videos not to play, content to flow in unexpected ways, and so on. Beyond that, our website relies on internet advertisements to pay the bills, so that you can read it for free. Server costs, site development and maintenance, research, and creating content all cost a lot of money, and without advertising revenue, our site may fade from existence. We are real people with jobs and families. If you enjoy the content on this website, which is supported by advertisements, please disable your Ad Blocker. Thanks!

The message says that “our website relies on internet advertisements” to cover the expenses of “server costs, site development and maintenance, research, and creating content”. Clash Daily explicitly say that advertising revenue allows them to create disinformation. I then decided to try to find if there were any clues in the source code about who provides this advertising to Clash Daily.

This is the asynchronous code for AdSense, Google’s online advertising service. Despite the Google News Initiative to help combat fake news and to “ensure that quality news content is recognized across our platforms”, Google provides the revenue that keeps Clash Daily, a fake news website, financially viable. I then wondered if this was coincidence and that maybe only a few websites happen to use AdSense. I decided to expand my search and ended up looking through the code for 18 different fake news websites.

This sites I’ve used are a mixture of clickbait attention-grabbing news and more ‘traditional’ fake news/disinformation. I’ve used both types of news in order to see if there are patterns of features being used across different types of fake news. As with the majority of fake news sites, few of these sites exclusively peddle fake content and they tend to be a mix of factual reporting and deceptive reporting in order to appear as legitimate factual sources. It is also important to note that some sites in the sample above have changed recently, whether they have become inactive, such as with nationonenews.com, or if they have even rebranded, such as yournewswire.com now redirecting to newspunch.com. I searched the sites above for AdSense and also read the full source code for each site to see if there are other services that fake news sites commonly use besides advertising. Here’s what I found (click on the table to view the full sized image):

The table above shows the 18 sites with a check next to them for 14 different features found across the sites, ranging from AdSense to PayPal. The sites with the most features are http://dailyheadlines.com/ (10) http://dailyheadlines.net/ (9) and http://thecommonsenseshow.com/ (8). The most used services are Google Fonts (18), WordPress (17) and Google Analytics (15).

The Services

This is where this long form blog post differs from the short-form version that I have also posted. I’ll now move onto a discussion of the features in the above table that will aim to give a description of each the service alongside the implications for the organisation that provides these services.

Google Services

AdSense

AdSense is the arm of Google that deals with online advertising. AdSense “provides a way for website owners to earn money from their online content” (Google, 2018) and website owners who sign up to AdSense will have ads provided by Google placed on their site that will generate revenue based on number of views and number of clicks. Google retains 32% of advertising revenue for ads placed on websites and in Q2 of 2018, advertising accounted for 86% of Google/Alphabet’s revenue with a 24% increase in revenue compared to the previous quarter. In other words: advertising is big money for Google. Here are some the ads on one of the sites, dailyheadlines.net: The use of ads on fake news sites isn’t just an issue for Google but it’s also an issue for the third parties who have trusted Google to publish their ads. Businesses who pride themselves on their image are being promoted on fake news websites unknowingly. The National Trust ad above appeared above an article titled “THANKS OBAMA! ISIS Makes TERRIFYING Statement About The SECRET Army That’s Already Here In Our Midst”. The article, that draws on ‘reporting’ from Info Wars, alludes to the notion of a ‘deep state’. This a common conspiracy that claims that while Obama was in office he established an army of intelligence officers to disrupt later governments and continue his policies. The National Trust is a registered charity and the most recent edition of their governance handbook (5^th ed, December 2016) says:

“[The Trust] Ensure accountability by: acting in accordance with the Nolan Committee’s Seven Principles of Public Life – selflessness, integrity, objectivity, accountability, openness, honesty and leadership”

Of these principles, “integrity” seems the most compromised as the National Trust are not only appearing alongside fake news, but indirectly funding it through paying for ads displayed on fake news sites.

reCAPTCHA

reCAPTCHA is a verification tool that protects websites from non-human users such as automated malicious software (bots). It usually involves either a challenge-response task where the user has to type two words on screen (v1) or tick a box next to the phrase “I’m not a robot” (v2) which will either allow the user to progress or present them with a v1-style visual identification task. 8 websites in this sample use reCPATCHA. Across all the sites, reCAPTCHA is used in the contact forms in the ‘contact us’ section. They’re used here so the website can’t be overloaded or incessantly spammed with messages from bots.

Fonts

Google Fonts is Google’s catalogue of over 900 fonts that are open source and free to use. Google Fonts can be embedded in WordPress sites using a Google Fonts plugin and APIs are available for CSS and Android implementation. Google write:

We believe the best way to bring personality and performance to websites and products is through great design and technology. […] All the fonts in our catalogue are free and open source, making beautiful type accessible to anyone for any project.” (Google, 2018).

Google fonts cache cross-site meaning that users only need to load a font once from any source and subsequent loading times will be decreased. As a result, Google Fonts offers websites improved performance and quicker loading times compared to other services. As Google note, their fonts are free to use and therefore “accessible to anyone for any project”, including fake news websites.

Design and how a website appears is important in whether a fake news story will be successful in its deception. Fake news websites function by appearing to be authentic members of the press and therefore a trustworthy news source. A site using professionally designed fonts that also load instantly will help contribute to the authenticity of the site; I’m not claiming here that fonts make a website seem legitimate, but they contribute to the perceived legitimacy of a website.

Google Tag Services

Used mainly for advertising, Google Tag Services sends information about the user to advertisers which is then used to create cookies. Tags can collect a variety of information including the content a user views, the links they click on, their IP address, type of browser, and the time they spend on each web page. Google simply defines a tag as a “snippet of JavaScript that sends information to a third party” (Google, 2018) and Tag Manager allows website operators to manage tags without having to edit the code themselves, thus opening up the use of tags to anyone (not just those who can code). This quote from a software engineer at Airbnb shows how effective Tag Manager is. However, when these tools are exploited by fake news producers it is equally easy for them to streamline their fake news services. The result is that sites can target both adverts and content they produce at certain users increasing the chances of a successful deception as they now know the audience they are pitching to; meanwhile, the ads generate revenue for the site. Targeting content at specific users is mentioned again later while looking at Roost and Criteo.

Google Analytics

Google Analytics allows website operators to see how many and how often people visit your website so they can optimise their content. Analytics records attributes as well as activities and can record country, age, gender, etc. of the web user. Websites use Google Analytics to analyse their web traffic to see what kind of people are visiting their websites. For example, if you discover that the majority of your visitors are 50-65 and they visit between 1700-1900, you can tailor your type of content as well as when you post it to capitalise on this information. Google notes that Analytics will help the user to “understand your site and app users to better evaluate the performance of your marketing, content, products, and more” (Google, 2018).

While Analytics helps countless companies such as independent businesses, start-ups, charities and other organisations, it is being exploited by fake news producers. Analytics shows you who is visiting your site and why, giving information about your readers and customer demographics. With this information, fake news sites can then tailor their content to the age and background of their known audiences. As fake news tends to prey on people’s irrationalities and emotions, using this data to inform the kind of fake stories you write can have a real impact on successful deception and potential further propagation of the story across social media; streamlining your website from appealing to the few to appealing to the many can be the difference between a single deception and mass deception.

Google Verified Site

“Verification is the process of proving that you own the site or app that you claim to own. We need to confirm ownership because once you are verified for a site or app you have access to its private Google Search data, and can affect how Google Search crawls it.” (Google, 2018).

Google allows website operators to verify their site with Google which consequently allows access to the Google Search Console via ‘Webmaster Tools’. These tools help give you not just general SEO but a Google-optimised site and helps increase Google search console visibility. Using Webmaster Tools, people who are unfamiliar with your site are more likely to find it in Google’s search console. This implicates Google in promoting fake news. If Google makes tools available for (fake news) websites that mean they can reach a wider audience and appear higher up in Google search results, then they are actively contributing to the spread of disinformation. There is also an issue of the authority that Google brings with it – if a site appears high up on Google’s search console, it is conferred an authority that makes it seem to be a reliable source. Consider this tweet:

In the tweet above from July 2018, the user refers to “Pages of evidence on google [sic]” and this demonstrates the legitimacy Google bestows on a source. By simply appearing in a Google search you can be considered ‘evidence’. As a result, Google providing Webmaster Tools to fake news sites means more people are potentially falling victim to fake news stories.

Overview: Google

There’s a fundamental dilemma here – all the tools that Google offer are great. Google boast all their tools are simple to use and freely available and they make huge, positive difference to all types of companies and organisations over the world. However, people are exploiting these services to disseminate and make a profit from fake news. This whole discussion boils down to one point: Google needs to do more to limit fake news sites using their services to help their disinformation operations. Fake news has proven to be deadly and by supporting fake news sites Google is complicit in these activities. A position of ‘ignorance is bliss’ can no longer be taken and Google needs to act in line with their own beliefs and ethics to cut off their services from producers of disinformation. Google’s role in helping fake news flourish is made more problematic by the existence of the Google Digital News Initiative. Launched in March 2018, the initiative states:

Google cares deeply about journalism. We believe in spreading knowledge to make life better for everyone. It’s at the heart of Google’s mission. It’s the mission of publishers and journalists.

Google goes on to state that “Quality journalism matters” but seem to address the potential criticism of supporting varied sources by saying “The digital news ecosystem should remain open”. The question is: does Google believe an open digital news ecosystem includes the availability of fake news?

Site Management

This section of the blog labelled ‘site management’ will look at services that operate at the top-level of a website. It contains only two services: WordPress, how sites are hosted; and Schema, a type of webpage classification. It’s a very simple section in a way as WordPress is used by all but one of the sites in my sample and Schema is used by all but four. It differs from the next section, ‘Site Optimisation’, in that WordPress and Schema are quite fundamental top-level services whereas the features found in the next section are optional features.

WordPress

WordPress comprises two distinct services: wordpress.org and wordpress.com. The former is the original WordPress site that currently accounts for 28% of the web while the latter is the full hosted version that means people can “start a blog or build a website in seconds without any technical knowledge” (WordPress, 2018). The WordPress network is, in simple terms, gargantuan:

Overall, the WordPress.com network welcomes more than 409 million people viewing more than 15.5 billion pages each month. Our users publish about 41.7 million new posts and leave 60.5 million new comments each month (WordPress, 2018)

WordPress is used by Yelp, The New Yorker, BBC America, Etsy, UPS and even Lancaster University (including this blog). WordPress has the power, if not the responsibility, to shut down websites that spread everything from political vitriol to racist and anti-Semitic views. All of these sites have their own .com or .net domain meaning they are paying for a premium plan with WordPress, which means WordPress is directly generating revenue from fake news websites. The story below is from yournewswire.com, a site hosted on WordPress:

The story above claims that a sex tape of Hillary Clinton and a child will prove Clinton’s part in operating a secretive Washington D.C. “pedophile ring”. The story above is similar to others published before, during and after the 2016 presidential election that ultimately led to Edgar Welch from Salisbury, North Carolina opening fire at a restaurant in D.C. alleged to be the hub of the paedophile network. This theory has been disproved by Fox News, Politifact, the Washington Post, CNN and various others. However, a paid-for WordPress site is still actively promoting the theory. While it is impossible (or rather implausible) for WordPress and other sites to track and limit all fake news sites, yournewswire.com is arguably the most notorious fake news site online and WordPress is continually allowing it to publish potentially dangerous fabrications while financially gaining from the process.

Schema

Jointly operated by Google, Yahoo, Microsoft and Yandex, Schema is a “structured data vocabulary that defines entities, actions, and relationships on the Internet (webpages, emails, etc)” (Semrush, 2017). Schema is used to structure the internet and different sites, pages, etc. are given classifications depending on their function. There’s a range of tags such as ‘website’, ‘person’, ‘NewsArticle’ among others (Full tagset: https://schema.org/docs/full.html). Structured markup data (which is what Schema helps do) can improve a website’s search engine in three ways:

1. Structured data helps Google algorithms to better understand and index your data.
2. Structured data imposes a structure on your websites architecture that makes it more efficient and therefore more search engine friendly
3. Structured data is easier to link automatically through different services (search engines; social media; other online platforms)

Schema helps the SEO of a site and therefore makes it easier for a site to be found on social media. Consequently, Schema helps increase the visibility of fake news. However, one thing to note here is that Schema is an undiscerning tool: it’s mission statement is simply to structure the internet to make it a more accessible and efficient however, as with Google, this tool is being exploited to deceive internet users. Point (3) above is important as it means that a search engine query won’t just give links leading directly to the original site (i.e. the fake news website itself), but will also provide results on other websites such as Facebook, Twitter, Wikipedia, other fake news sites etc.

Site Optimisation

Roost

Roost is used to create push notifications to display short messages to viewers of a website. Push notifications normally require the permission of the user, such as here where I visited the Roost homepage on Google Chrome:

Web-based push notifications are delivered to a user’s desktop as long as they have a browser open, irrespective of whether the original site sending the notifications is open. The reason so many sites use these notification is to essentially combat the notion of “out of sight out of mind”; if you can alert a user to new/relevant content even when they aren’t on your website, you will garner more views from people as they click on the notifications and are redirected to your site. Roost, however, does not simply display generic news alerts or messages to users.

Roost uses ‘geo targeting’ and ‘advanced segmentation’ that “segment site subscribers into groups based on common interest, content category, or opt-in page” (Roost, 2018). This means Roost can send highly targeted and personalised messages to users that are shaped on the content they have already viewed as well as other information available stored in cookies, such as other sites they have viewed, their location, their common browsing hours etc. The issue with this is it enables a system of positive feedback creating an echo chamber: users view content that they agree with/appeases their views, as a result they are then shown more content based on these viewing patterns, and this cycle repeats so users are only shown content that pertains to their beliefs/views. The result is the aforementioned concept of an echo chamber (a phrase bandied around often in fake news research), the concept that people seek out (or are provided with) information that cements their pre-existing beliefs and biases which can lead to increased sociopolitical polarisation. An example of this is the case of Darren Osborne.

Osborne drove a van into a mosque near Finsbury Park in London on Ramadan, killing one person and hospitalising twelve others. Scotland Yard accessed Osborne’s browser history and found he became radicalised in just three weeks, releasing in a statement that:

It was clear that within the space of only a few weeks Osborne had developed a warped and twisted view.

Osborne cocooned himself in far-right wing sources on Twitter and Facebook and became radicalised. He repeatedly accessed sources that further perpetuated his views and this ultimately led him to kill someone. This is an extreme case of radicalisation but it demonstrates the danger of reinforcing and sustaining biased politically-charged viewpoints. Imagine someone has personal views such as wanting to decrease immigration (a view they are entitled to hold), they hold this view and do not act upon it; however, when they are constantly shown highly-targeted and divisive information straight to their desktop or phone, this view intensifies and extremifies and they risk going down the route of intensifying their views to the points of radicalisation. I’m not claiming here that Roost (or other services) are a direct route to extremism or radicalisation – instead I’m saying that features such as these are a part of the mechanism that contribute to the development of extremist views. Features such as Roost on sites peddling misinformation help foster social and political polarisation, but they are not in themselves a silver bullet that creates divides.

Yoast

Yoast is a WordPress plugin for SEO that, similar to Google Site Verification and Schema, helps your site be found on search engines such as Google and Bing. Yoast “helps you to optimize your website for search engines” (Yoast, 2018) and allows you to optimise features of a WP site, such as: title optimisation; permalink structure; code refining; site structure; off-site SEO (social media); etc. Yoast is very commonly used and is a great tool for businesses to create awareness of their brand and to bring in new customers. However, as is a repeating pattern, this tool is being exploited.

Imagine this seconario: a U.S. citizen holds the view that the country should shut down its borders and are concerned about the migrant caravan heading from Honduras through Mexico to the US border. On the back of remarks from the US president that Democrats are supporting the caravan of the migrants, if they were to Google something like “migrant caravan dangerous”, Yoast, as an SEO tool, would help direct them to articles such as this. This article on YourNewsWire (a site that uses Yoast) falsely claims that philanthropist George Soros is funding the migrant caravan and that those in the caravan are paid activists. What might start as a political view is compounded by an array of false information that is made easy to access by SEO tools such as Yoast. YourNewsWire disseminates anti-semitic and racist views yet Yoast, a company that “want[s] people to focus on producing quality content” (Yoast, 2018), is allowing them to be a customer and are thus helping propagate extremist views online.

Criteo

Similar to Google’s in-house service Ad Sense, Criteo is a tool that tailors adverts for people who have visited your site before in order to show them more apposite ads. It uses a form of advertising knows as ‘retargeting’:

“Retargeting is a form of advertising that helps brands re-engage users who’ve left their websites before making a purchase. There are several types of retargeting [including] cookie-based retargeting, or the one retail marketers are most likely to be involved in on a day-to-day basis.” (Criteo, 2018).

The purpose of Criteo is to encourage users to re-engage with brands. Criteo works by dropping an anonymous and undetectable browser cookie on users so that when they revisit your site, you can retarget them and show them ads based on their previous behaviour. Crieto themselves concede that retargeting “sounds a bit creepy and geeky”.

Criteo’s ‘About’ page states they are a “a transformative company committed to excellence” but this form of advertising is seen as an inherently negative practice in the industry. Following Apple’s announcement that it will seek to limit tracking on its desktop version of Safari using Intelligent Tracking Prevention (ITP), Criteo announced that they expect their revenue to drop by around 22% on the back of Apple’s ITP efforts. That one of the world’s biggest tech manufacturers is seeking to limit this kind of tracking shows what a considerable issue it is. Criteo marks an interesting case of services used by fake news sites. Until now what we have seen is tools being exploited to benefit fake news producers, whereas Criteo appears be a case of fake news sites using methods/tools that are already frowned upon in the tech community.

Mailchimp

Mailchimp is integrated into a website to allow users to subscribe to marketing emails. These normally take the form of newsletters or marketing campaigns from the website to, in the case of fake news sites, update subscribers about the latest stories or to sell products. Mailchimp is a form of marketing automation and according to iDataLabs they are used by over 350,000 companies worldwide and have a market share of 43.65% in the US market. MailCcimp is overwhelmingly favoured by small companies, with companies of 1-10 or 10-50 employees (58% and 29%, respectively) comprising the majority of their customer base.

Mailchimp say that they “provide sophisticated tools anyone can use to build their brands and sell more stuff” (Rocket Science Group, 2018) and help companies to establish their brand presence and subsequently to generate revenue. Mailchimp is a great tool for small companies – it has three price tiers, of which one is free, and can help businesses build up brand loyalty by creating email campaigns, designing mailing list landing pages, and operating fully automated services such as welcome emails and order notifications. The service Mailchimp provides can be seen as combining the monetary function of AdSense with the awareness function of Roost in that subscribers are sent newsletters and updates but also sent emails if they order a product.

Although not in this dataset, a site peddling disinformation that exemplifies how fake news sites sell products is Alex Jones’ infowars.com. Late night host John Oliver quoted a former employee of Jones’ as saying that Jones “can sell 500 supplements in an hour … it’s like QVC for conspiracy” (source). The homepage of Info Wars features ads such as these:

Although Info Wars’ store is not powered by Mailchimp, the site acts as a perfect example of how fake news providers can generate revenue. According to the New York Times, Jones’ operations generate over $20million a year and “most of his revenue [comes] from the sale of products like supplements” (New York Times, 2018). Operating a large website that is regularly updated with content can be very expensive and fake news sites can use services such as Mailchimp to help generate revenue via product sales, essentially allowing them to continue posting intentionally false news content.

Payment

PayPal

Of all the services discussed in this post so far, I imagine PayPal is the only one that requires no introduction. PayPal is a ubiquitous online payment platform used by 7 million businesses in 202 countries and 25 currencies, including: Macy’s, Home Depot, Abercrombie & Fitch, Ocado, Dominoes pizza, Dell, and even the Church of England. Formerly owned by eBay but now a separate entity, PayPal has “received more than 20 awards for excellence from the internet industry and the business community” (eBay, 2018) and in the year 2017 generated revenue of 13.09 billion USD, an increase of just under 21% on the previous year. PayPal appears in the code for three sites: yournewswire.com, thecommonsenseshow.com and the now defunct gotnews.com. However, YourNewsWire does not have a shop section nor any other section where payment is taken, so it is likely that this is redundant left over code from a previous iteration of YourNewsWire. For the Common Sense Show, PayPal is used on the section of their site titled “Help Support Our Research”. This is a donation page where readers can donate money to help support the page. The page reads:

The Common Sense Show works hard to keep the public informed about a variety of topics.  Please consider helping us to offset our investigative, research and broadcasting costs. We have provided two convenient ways to contribute to our efforts.

This is similar to the message about ad blockers higher up on clashdaily.com where there is a plea in order to support the site’s journalistic intentions. Using PayPal, the Common Sense Show can receive cash donations to fund its writers. This also presents another issue: if PayPal were to rescind their platform so TCSS couldn’t use it, they would simply find an alternative and use them (such as Stripe, discussed in the next section). This shows the endemic issue of supporting fake news, an inter-industry wide initiative is required to inform companies they are supporting fake news so they can withdraw their features. While this is unlikely to ever happen, if very large companies like Google, WordPress and PayPal withdraw their support of fake news sites, it’s likely others will follow. Those who don’t follow suit will then be characterised for their support of sites with extremist views, damaging their reputations and their relationships with legitimate companies (no one wants to use a company known for supporting racism etc).

Stripe

Similar to PayPal, Stripe is an online payment platform serving some of the most well known companies in the world, including: Amazon; Facebook; Deliveroo; Target; Uber; TED; ASOS; Missguided; Under Armour; Lyft; laundrapp; National Geographic; UNICEF; Oxfam; NPR; Comic Relief; and more. The benefit of Stripe over PayPal is that it is very developer friendly. Stripe can be fully integrated into a system to the extent that Stripe is hosting the transactions carried out on a site but their branding or name does not appear anywhere on the site. This arguably makes Stripe more omnipresent than PayPal as it can be powering a transaction without you realising. Stripe comment that “the best platforms own their user experience from end to end, including payments” (Stripe, 2018) and allow companies to appear to be handling their payment independently when in fact they are supported by Stripe. This also shows the utility of looking through a website’s code – it can reveal the services such as Stripe that operate behind the scenes. As with PayPal, there is an issue here of if Stripe banned fake news sites, then what is to stop the site from using a competitor in what is a highly competitive market.

Conclusion

Part of the purpose of this blog post was to implicate companies supporting fake news websites by showing how their core values are incompatible with their support of disinformation; this is why throughout the post I’ve often quoted a company’s values (usually from their ‘About’ page) and contrasted the statements with their practices of supporting content that fosters hatred and encourages prejudice. It could easily be argued that fake news sites don’t pose that much of a risk and that people don’t always believe fake news stories. However, sites such as the sample used in this post undoubtedly foster hatred.

Here’s an example of the comments on a 2017 YourNewsWire story about a Hillary Clinton “Pedophile Sex Tape” that perpetuates the notorious pizza gate conspiracy. It took less than a day for people to post anti-Semitic, anti-Muslim and racist comments below the article despite the original article not addressing these themes (I have blanked out racist terms):

Fake news websites do not exist within an entirely fake vacuum – a website’s content is just the surface and looking behind this surface can reveal a lot about how fake news sites function. Fake news articles and disinformation campaigns foster hatred and stoke political and social tensions and they are being propped up by companies and organisations that claim to inherently conflict with these practices.

What is being done about this?

As I mentioned early on in this post, during the process of writing this post a number of sites I used have shut down and return 404 error pages or redirect to the domain owner. It is unclear whether the sites were shut down by the website host (WordPress) or if it was voluntarily closed down, something that often happens when fake news sites gain notoriety (see the ‘Denver Guardian’).

A big component of this blog post is guilt by association – example ad 1 at the top of this post (the National Trust) demonstrates this well. The National Trust, a conservation charity, is paying money in advertising costs to unknowingly appear above messages that are often overtly racist and anti-Semitic. While researching this blog post, I found examples of the National Trust, the Natural History Museum, Jet-2 Holidays, Honda, Sky Betting, and other well known British and global businesses advertising on fake news websites. It is unlikely that any of these organisations, especially those that are charities, would be happy with their indirect supporting of these websites. What could an outcome of this be? Well, for many organisations their hands are tied; they could potentially withdraw their advertising from Google on moral grounds, but this would considerably affect their businesses. Another option would be to request that their ads don’t appear on these kind of websites. This is where the British organisation Factmata comes in.

Factmata is developing a quality-based metric to “help advertisers avoid placing advertising on fake news, hate speech, and extremist content” (Factmata, 2018). What this means is that Factmata plans to assign quality scores to websites so advertising companies and advertising purchasers can be aware of where their ads are being placed. Dhruv Ghulati, Founder and CEO of Factmata, said the following in an interview earlier this year:

We’re not in the business of trying to remove content or remove opinion from the internet. We’re here to flag potentially sensitive or biased content by providing brands and media agencies with a quality score which indicates the likelihood that a page contains misinformation. Media buyers can then use that score to decide whether they want to take the risk of advertising on that site or not.

Measures to tackle fake news are often attacked on the basis of censorship, however the Factmata method does not propose removing content, but simply starving it of financial support. They can be seen as indirectly shutting down fake news sites by removing the services of legitimate businesses. This is an especially important issue in the U.S. where the First Amendment right to free speech is often used as a shield for not directly closing down fake news websites. If Factmata are successful in their mission to help advertisers avoid sites peddling extremist views, it could be a much more crippling blow to fake news than any literacy education, awareness campaigns or even direct censorship of fake news could achieve.

To view the short-form version of this post, click here.