YouTube claims that between January and March this year, it removed over 8,20,000 videos in India alone that were violating its community guidelines. It also updated its policies last year to limit hate speech violations that could be attacking a particular caste.
YouTube is planning to add Hindi support to its mechanism that reduces the recommendations and visibility of content that could spread conspiracy theories, an executive said. The change is specifically designed for the India market. It will be an extension to the company’s move that allowed cutting down the video recommendations that could misinform viewers in harmful ways, which it calls “borderline” content. YouTube is essentially trying to address user concerns by limiting content spreading conspiracies and misinformation on its platform.
Chief Product Officer Neal Mohan told Gadgets 360 over an email that YouTube had already started reducing recommendations of borderline content in English and was planning to launch its Hindi support this quarter.
In January last year, YouTube started to reduce content recommendations spreading conspiracy theories, such as videos promoting a miracle cure for a serious illness or claiming the earth is flat. However, that development was initially limited to the US market, though it expanded to Australia, Japan, and New Zealand later on. The update also reached initial non-English language markets including Brazil, France, Germany, Mexico, and Spain in December. Additionally, YouTube claimed that it saw a 70 per cent average drop in the watch time of such content coming from non-subscribed recommendations in the US.
Mohan told the reporters during a video conference earlier on Friday that YouTube had made over 30 changes to its recommendation algorithms over the last year or so to reduce borderline content that isn’t exactly against its policies but “might be harmful misinformation in some other way” and could impact the lives of viewers.
Language is a key barrier for YouTube especially in India to limit videos that are created to attract clicks and views on the platform and plot misinformation. However, Mohan, while responding to a question raised by Gadgets 360 during the conference, said that YouTube’s content rating team already covers many languages throughout the country.
“One of my big goals is to continue to expand and deepen that language coverage for all our raters,” he said. “We also work very closely on our content policies with organisations in India to make sure that our policies and our enforcement guidelines reflect the specific conditions in India so that we can quickly act on content that might be violated on our platform.”
Ramped up moderators hiring
Mohan told Gadgets 360 in the follow-up email that his team had ramped up its hiring significantly, bringing the number of people across Google focussed on moderation issues to 10,000 last year. Software-level changes are also taking place in the background to reduce deleterious content automatically to some extent.
There has been a serious criticism around how YouTube polices content on the platform. Although its guidelines and policies don’t explicitly allow objectionable content, the company has been alleged for going ahead with the videos just for sake of favouring some of its creators and in a handful of cases may be just due to failure to enforce its rules and regulations.
Nonetheless, Mohan said that the moderators work in a “24-seven, ‘follow the sun’ model” to make sure that the content is at par with its standards. He also underlined that the policy enforcing teams at YouTube sit across time zones and cover different languages including most of the “commonly spoken Indian languages” as well as have different areas of expertise.
