Meta said on Tuesday it has strengthened default protections and recalibrated Instagram’s content recommendation systems in India to reduce the visibility of sexually suggestive material and content depicting violence. In a Facebook post titled “An Update on the Content We Recommend on Instagram in India”, the company said the work was carried out over the past year and guided by user feedback on what they want from recommendations. Meta said it is deploying more capable artificial intelligence models that screen content before it is recommended, rather than waiting for a report or hide action, and that these systems have been tuned specifically for India.
The changes follow scrutiny of Meta’s content moderation systems. In July, the Ministry of Electronics and Information Technology (MeitY) issued a notice to Meta over alleged child sexual abuse material (CSAM) promoted on Instagram and ordered that it be disabled, and summoned the company’s global executives to India. The scrutiny intensified after a BBC Eye investigation reported observing around 30 unique advertisements on its platforms allegedly promoting CSAM. On October 10, MeitY sent Meta an advisory asking it to ensure recommendation systems do not show only a single or limited set of narratives and to provide diverse information, and to move faster on spotting and escalating posts amplified through recommendations that could cause confusion, violence or public disorder, as first reported by HT.
According to Meta, people in India now report recommended content about half as often as they did a year ago. The company did not share numbers or specify when the India-specific changes went live. It pointed to user controls, including Sensitive Content Control, which allows users to opt for less sensitive material across Search, Reels, Explore, suggested accounts and in-Feed recommendations. Instagram said its systems in India already exclude the most mature content from recommendations, with the setting allowing users to further limit exposure to other sensitive material.
Users can also choose to see less political content from accounts they do not follow, flag posts through the “Not Interested” option, and filter words, phrases and emojis using the Hidden Words feature. The platform’s blocking function prevents specific accounts from finding profiles, viewing posts or sending messages, including through new accounts they may create. Teen Accounts, launched in India last year, place teenagers in a 13+ content setting by default. Meta said content that breaks its Community Standards is removed, while content unlawful in a particular country is restricted there, citing its global enforcement report and the monthly transparency report it publishes under the IT Rules, 2021. HT reached out to Meta and MeitY for comments but did not get them immediately.
