Facebook and Google have recently being getting flack from users for hosting fake news stories. As Jack Nicas and Deepa Seethataman report, “Google and Facebook Take Aim at Fake-News Sites”

Facebook Inc. and Alphabet Inc.’s Google announced steps to prevent fake-news websites from generating revenue through their ad-selling services, signs that technology companies are moving to address a growing controversy about misinformation on the internet.

The Fake News Game is illustrated in Figure 1, and players’ objectives, actions, and tensions are described briefly below.

Figure 1

Advertisers

Objective: Maximize User ad clicks

Action: Choose sites, including News Sites and Platforms, to host ads

Given: Existence of both real and fake news and information

News Sites

Objectives:

• Generate news stories to attract Users

•Maximize revenues from User clicks on hosted ads.

Actions: Generate news stories from available information to attract both Users and Platforms to both read and repost stories

Given:

• Existence of both real and fake news and information

• Costs (e.g., time) of verifying news and information for legitimacy

Tensions:

Real News Sites: Want news to be

• Interesting so as to attract Users

• Real so as to maintain reputation for reporting real news

Fake News Sites: Want news to be

• Interesting so as to attract Users

• Similar to Real News so as to avoid being filtered out

Platforms

Objectives:

• Host news stories and information to attract Users

• Maximize revenues from User clicks on hosted ads.

Actions:

• Host news stories to attract both Users and Platforms to read and repost stories

• Design algorithms to filter out fake news and information

Given:

• Existence of both real and fake news and information

• Costs (e.g., time) of verifying news and information for legitimacy

Tensions:  Design algorithms, given

• If weed out too much news (i.e., falsely exclude real news), then lose ad revenues

• If weed out too little news (i.e., falsely include fake news), then lose/anger Users

Users

Objectives:

• Visit websites that provide interesting news and information

• Generate reputation for passing along (reposting) real and interesting news and information

Actions:

• Choose websites to visit

• Repost interesting news and information

Given:  

• Existence of both real and fake news and information

• Costs (e.g., time) of verifying news and information for legitimacy

Tensions: Choose news and information to repost, given

• If repost too little news and information, then lose reputation

• If repost too much fake news and information, then lose reputation

Observations

• The more difficult it is to verify (and weed out) news and information for accuracy and legitimacy, the more fake news and information will appear.

• The more fake news there is that is reported or posted, the more difficult (more costly) it is to verify the accuracy of any given piece of news and information. That is, there is a feedback loop here.

• The lower is the penetration of fake news and information, the greater is the tendency for Users to assume news and information is generally true.

• However, once a minimum threshold of misinformation has been achieved, Users can no longer assume that news and information is generally true. I think we’ve almost achieved this tipping point.