Abstract

Digital platforms experience new levels of complexity as their reach and interactions increase. Every day, thousands of pieces of content are classified as unsuitable for permanence on the web. The balance between forms of control and the actions of users has led Meta’s administration toward the implementation of the Oversight board (Mob), a guarantor of freedom of expression and online safety that can provide independent judgment on compliance with the Community standards. Based upon the guidance given by the regulations, content moderation employs two approaches: horizontal, with user reports; vertical, through automated detection driven by Artificial intelligence and the supervision of reviewers, who are employed by the company. While content moderator is just one of the temporary jobs in the gig economy, Ai guarantees efficient control, identifying content classifiable as spam, but it risks encroaching on the field of decision making in certain cases requiring protection such as freedom of expression. Therefore, the Oversight board has a decisive role in determining «what to remove, what to leave, and why», although it does not directly affect the algorithms and market business strategies. Beginning with decisions published from 2020 to 2022, this article profiles the risks citizens face in exercising their rights in relation to the code and the predictive ability of algorithms to implement qualitative discriminations.