Experimenting

This week I continued my work on the Democratizing Big Tech! We are addressing the effects of hate speech and anti-migrant speech on social media sites.  It was quite a coincidence that this was the same topic from journal write on but just on the global stage.  I started by looking at policies and frameworks that other countries use.  But first, it was crucial that I understood the terminology as it is important to recognize the difference between disinformation and misinformation. The International Center for Not-for-Profit Law (ICNL), an international organization working to improve the legal foundation for philanthropy, civic engagement, and civil society, defines disinformation as “the intentional dissemination of misleading and distribution of wrongful information.” Disinformation may also include truthful components or “half-truths” which makes it difficult for users to recognize the false statements. The ICNL defines misinformation as the unintentional dissemination of misleading information, essentially it is “the act of people mistakenly or in error deceiving others, while the person sharing the false information is unaware of it. 

Surveillance advertising is defined as the practice of displaying various advertisements to different customers based on assumptions made about their interests, demographics, and other details that have been obtained through monitoring their behavior.  To gain this information, the companies track their consumers by creating profiles of each consumer and matches advertisements. To build these profiles, Ad Tech companies, monitor user behavior on platforms and then auction off the data to sell the right to advertise to that person.  In this way, platforms such as Google and Facebook can profit off users and sell their data.   While this practice is not directly linked to the spread of misinformation and hate, it is possible that it can contribute to the spread of such content.  When AdTech companies create profiles of consumers, they learn about the users’ interests, demographics and online behavior.  It is easy to then manipulate that information by creating tailored messages to reasonable to biases, fears, or prejudices of specific groups of users allowing the harmful content to spread quicker and amplify. 

 

In the United States of America (USA), Congress passed 47 U.S.C. § 230 which permits tech companies to moderate the platforms without being legally liable for the content which they host. Such legislation is to blame for the lack of accountability in the Big Tech industry in the USA. Social media platforms are said to curate content rather than serving as a simple container for other people's content. This is conveyed through the mission statements of the platforms, YouTube aims to be family-friendly, Twitter is a place for free speech, while Facebook aims to foster authenticity and safety. In order to do this, social media sites curate their platforms to fit the environment they intend to foster.  This is done by these platforms stating that they will monitor and remove content that goes against their message.

 

In order to determine if the platforms are actually doing as they say we teamed up with Global Witness a non-profit from Denmark that has done similar projects in the past.  In the meeting with them, we created a plan for how to embark on the project.  Each candidate attorney had a platform they were in charge of Facebook, Tiktok, and Twitter.  To determine if they were monitoring content that included hate speech, we had to make a separate email account and make sure that the account was only used on one device that did not have other accounts logged in so as not to raise suspicion. Then the hate speech content would be sent in as ads, 1. so that they wouldn’t actually be published and 2. Ads have to be approved so it would ensure that the hate speech had been viewed by a member of the staff.  The other issue was that English was not the only language in South Africa so it was important to establish whether these platforms were removing hate speech that was written in these common languages.

 

Next is the taxing part of creating around 70 ads per platform in order to conduct the study. I am excited to a part of this experiment and cannot wait to see the results!