Meta’s AI moderation and free speech: Ongoing challenges in the Global South
Bibliographic record
Abstract
Abstract This study investigates the discriminatory impact of artificial intelligence (AI)-driven content moderation on social media platforms (SMPs), particularly in the Global South, where cultural and linguistic diversity often clash with the Western-centric AI frameworks. Platforms like Meta increasingly rely on AI algorithms to moderate vast amounts of content, but research shows that these algorithms disproportionately restrict free expression in the Global South (European Union Agency for Fundamental Rights, 2023; De Gregorio & Stremlau, 2023). This results in “over removal” – censorship of lawful content – and “slow removal,” which fails to address harmful material, both of which perpetuate inequality and hinder free speech. Through a case study on Meta, this research examines how AI-based content moderation misunderstands local contexts and systematically marginalizes users. The contributing factors include limited financial investment, inadequate language training, and political and corporate biases. The imbalance reflects power asymmetries, as governments in the Global South lack influence over platform policies. This study uses a human rights perspective to explore solutions through multistakeholder engagement, advocating for collaboration among tech companies, governments, and civil society to reform AI governance. Ultimately, it aims to inform regulatory frameworks that ensure fairer, more inclusive content moderation and protect free expression for a globally equitable digital landscape.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.016 | 0.024 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.007 | 0.016 |
| Scholarly communication | 0.010 | 0.014 |
| Open science | 0.001 | 0.011 |
| Research integrity | 0.002 | 0.004 |
| Insufficient payload (model declined to judge) | 0.011 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".