Warning. The editorial policies of Facebook allow for the promulgation of fake or deliberately inaccurate information presented as "news". If you have your own account on Facebook, you may be served with fake or deliberately inaccurate information presented as news.

Continue to Facebook

Here is an example (screenshot from Facebook, 12/9/2016):

Dear Facebook,
Who paid for this?

Making an analogy with print media, it reads kind of like a National Enquirer headline. But, it was shown on a Facebook page as part of somebody's paid make-fun-of-the-Clintons advertising campaign.

This particular example is obviously fake to many users, but not necessarily to all users. The point is this: Facebook users need to make the effort to identify each news-like item as fake, or not, because Facebook itself is not making the effort.

Update February 25, 2017:

I'm still clueless about why that particular vicious add was served to me. This article in Scout, The Rise of the Weaponized AI Propaganda Machine (Berit Anderson and Brett Horvath) suggests how it may have happened.
For Analytica, the feedback is instant and the response automated: Did this specific swing voter in Pennsylvania click on the ad attacking Clinton's negligence over her email server? Yes? Serve her more content that emphasizes failures of personal responsibility. No? The automated script will try a different headline, perhaps one that plays on a different personality trait -- say the voter's tendency to be agreeable toward authority figures. Perhaps: "Top Intelligence Officials Agree: Clinton's Emails Jeopardized National Security."

Much of this is done through Facebook dark posts, which are only visible to those being targeted.

Sounds like an evolution from Web 2.0 -- instead of suggesting vacations we might want to take in our search results, the bots are going to browbeat us with ads until they finds a wedge issue that will make us want to go vote for the candidate dear to the heart of the paying client.

Update August 4, 2017:

Here is a hypothesis about why that particular vicious add was served to me: One of Facebook's political advertisers had me down as a Bernie supporter, and they were trying to discourage me from voting by reminding me of the Clintons' past. Except that, it was December 9 (I'd already voted), and I wasn't actually a Bermie supporter.

Here is a great book review called by Sue Halpern in the New York Review of Books. She reviews Prototype Politics: Technology-Intensive Campaigning and the Data of Democracy by Daniel Kreiss and Hacking the Electorate: How Campaigns Perceive Voters by Eitan D. Hersh.

Facebook did turn out to be essential to Trump’s victory, but not in the way Grassegger, Krogerus, and Schwartz suggest. Though there is little doubt that Cambridge Analytica exploited members of the social network, Facebook’s real influence came from the campaign’s strategic and perfectly legal use of Facebook’s suite of marketing tools.


In the course of the 2016 election, the Trump campaign ended up relying on three voter databases: the one supplied by Cambridge Analytica, with its 5,000 data points on 220 million Americans including, according to its website, personality profiles on all of them; the RNC’s enhanced Voter Vault, which claims to have more than 300 terabytes of data, including 7,700,545,385 microtargeting data points on nearly 200 million voters; and its own custom-designed one, called Project Alamo, culled in part from the millions of small donors to the campaign and e-mail addresses gathered at rallies, from sales of campaign merchandise, and even from text messages sent to the campaign. Eventually, Project Alamo also came to include data from the other two databases.

A principal force behind these various strategies was Brad Parscale, who served as the digital director of the Trump campaign from the primaries through the general election and who in the late spring of 2016 hired Cambridge Analytica as part of this effort.


While it may not have created individual messages for every voter, the Trump campaign used Facebook’s vast reach, relatively low cost, and rapid turnaround to test tens of thousands and sometimes hundreds of thousands of different campaign ads. According to Issie Lapowsky of Wired, speaking with Gary Coby, director of advertising at the Republican National Committee and a member of Trump’s digital team:

    On any given day…the campaign was running 40,000 to 50,000 variants of its ads, testing how they performed in different formats, with subtitles and without, and static versus video, among other small differences. On the day of the third presidential debate in October, the team ran 175,000 variations. Coby calls this approach “A/B testing on steroids.”