Jornal O Globo - November 18, 2024
Social networks lack transparency in data provision and advertising display, study concludes
As dependence on large digital platforms grows, so does the urgency for greater transparency in their services. This is particularly true in Brazil, according to a recent study by NetLab, a laboratory linked to the School of Communication at the Federal University of Rio de Janeiro (UFRJ).
The study evaluated the platforms using two indexes. The first measured transparency in data provision, the second in advertising display. When subjected to the evaluation, the main platforms presented poor results. None reached the ideal level. In the Data Transparency Index, YouTube obtained the best score (63.2 points on a scale of 0 to 100), a level that was only satisfactory. Facebook (53.6) and Instagram (52.1) performed averagely. WhatsApp received a minimum score of 1.5 points, revealing a serious lack of transparency. In the Advertising Transparency Index, Meta — owner of Instagram, Facebook and WhatsApp — obtained the best score (49.8), although it was only fair. Next, Telegram (22.8), Linkedin (18.3) and Google (8.2) presented a precarious level.
One of the most significant concerns highlighted in the study is the difficulties in accessing information. NetLab researchers criticize the increasing restrictions on the platforms' Application Programming Interfaces (APIs), essential tools for data collection. They specifically mention the abrupt termination of Meta's CrowdTangle tool, which previously allowed access to information from Facebook and Instagram. The restriction, in addition to preventing independent research, allows the platforms to selectively release incomplete or inconsistent data. The researchers advocate criteria similar to those required by European law.
The study also criticizes the lack of transparency around content moderation practices. According to the researchers, details are lacking in the information provided in the platforms' transparency reports. While they do disclose the number of posts removed for violating the rules, they generally do not provide information about the specific types of violations or the criteria used to make removal decisions.