The non-profit news site hopes (The Markup(Through his new project called)Citizen Browser Project) In demystifying algorithms The YouTube And Facebook, explaining why everyone sees the platforms differently from the other.
Technology companies use algorithms to organize content, even though the way these algorithms work could have serious social and political consequences.
Julia Angwin says (Julia AngwinThe Markup editor: Social media platforms are the broadcast networks of the 21st century.
“These platforms identify news that the public consumes, using algorithms designed to maximize profits at the cost of truth and transparency,” she added.
The Markup has created a group of 1,200 people who are supposed to use a private browser when browsing Facebook and YouTube.
The browser sends real-time information about the content presented to participants – only the content provided, not the way users interact with it.
Presumably, over time there will be statistically valid data that provides insights into how Facebook and YouTube algorithms work.
“Browsing via private browser allows researchers at The Markup to make a connection between the personal characteristics or demographics of a particular user and the content being viewed,” Angwin says.
Browsing via the private browser also reveals to researchers the differences in the way the platforms present content to black users versus white users, or conservative users versus liberal users.
Angwin believes that browsers are showing whether some users are seeing more harmful content than others, such as disinformation about the coronavirus.
The browser data should also shed some light on how Facebook uses personal and demographic data to accurately target ads.
Facebook’s publicly available ads library provides a certain amount of analytics data that generally shows how the ads are targeted.
Angwin says the company does not provide the full picture of the user’s personal or demographic data that triggered the ad.
The Markup group begins using the browser within the next few weeks, and will continue using it at least until the January inauguration.
Angwin notes that the reality of the algorithm is not the way it was written but rather the results it produces.
She explained that major technology companies are increasingly using machine learning to suggest content, and in this case there is no real algorithm that can be seen, just a complex probability measurement system that is constantly interacting with new data.
The non-profit news site plans to publish its findings in a series of articles in the New York Times next year.