The Center for Democracy and Technology (CDT) has released a report titled "Moderating Tamil Content on Social Media," focusing on the challenges faced by Tamil speakers in online environments. The study highlights the complexities of content moderation for Tamil, a language spoken by over 80 million people worldwide, primarily in India and Sri Lanka.
According to the report, there is limited understanding of how major social media platforms moderate content in Tamil. This issue is part of a broader series examining content moderation within low-resource and indigenous languages in the Global South. The lack of high-quality datasets hinders effective automated moderation systems, as previously observed with Maghrebi Arabic and Kiswahili.
The research explores the experiences of Tamil users on popular social media platforms run by Western and Indian companies. It notes that Tamil speakers perceive poor content moderation amidst democratic backsliding and growing repression of speech in India and Sri Lanka. This perception creates an environment where users encounter both over-moderation and under-moderated feeds containing hate speech.
The study employed a mixed-method approach, including an online survey of 147 frequent social media users from India and Sri Lanka, 17 interviews with various stakeholders, and a roundtable discussion with experts. Findings indicate that while Western social media platforms are popular among Tamil speakers, local TikTok alternatives are gaining traction following India's TikTok ban.
Tech companies use global or localized approaches to moderate Tamil content. The global approach applies uniform policies worldwide without linguistic expertise considerations, whereas the local approach tailors policies specifically for Tamil language guidance. Some Indian companies adopt a hybrid method with occasional localized adjustments.
Tamil speakers face inconsistent moderation attributed to their primary language not being English. They experience both under-moderated environments filled with abuse in Tamil and perceived unfair over-moderation targeting their speech.
Political motivations are also a concern among respondents who believe that content removals are used to silence political discourse. Some suspect they experience "shadowbanning" when using specific words or symbols associated with the Tamil community.
Despite having an active computing community, investment in automated moderation for Tamil remains inadequate due to resource limitations faced by smaller social media companies.
For more detailed insights, readers can access the full report provided by CDT.