Facebook announced that it has brought the automatic captioning feature, which has been used on its platform for a while, to Instagram TV. It is stated that the feature, which can create automatic subtitles in 16 different languages for now, can be used in more countries and languages in the future.
Social media giant Facebook launched automatic captioning for Facebook Live and Workplace Live in March. Now expanding the scope of this feature, Facebook announced that the automatic captioning feature is also available for Instagram TV videos. The new feature of Instagram TV can be used in 16 different languages.
In a statement made by a Facebook spokesperson, it was announced that the feature is only in its initial stages and that many more languages will be supported in the coming periods. The feature powered by artificial intelligence seems to attract the attention of users who watch foreign language videos on Instagram TV.
The automatic captioning feature, now available on Instagram TV, is based on machine learning. Artificial intelligence, which has the ability to perceive the conversations in the video, can translate the perceived speech into specified languages and turn them into subtitles. Engineers state that they developed this artificial intelligence a little more before bringing it to Instagram TV and taught words such as “covid”, which have been frequently pronounced recently.
Facebook is one of those companies that has been caring about automatic captioning for a long time. In fact, the social media giant announced its first automatic captioning product for ads in 2016. By October of the same year, a free tool for translation for US English was released. However, to be honest, Facebook is not yet at the level it wants to create automatic captions.
This is how Instagram TV’s new feature looks
Facebook has been the focus of criticism in the past due to its automatic captioning feature. The reason for this was that some words in particular were translated in a rather ridiculous way. This situation posed serious problems, especially for hearing impaired users. Fortunately, Facebook is aware of this situation and is trying to find solutions.
In a technical article published last month, Facebook mentioned that they are developing a new speech recognition software framework. The officials stated that this software framework, called “wav2vec 2.0”, was very successful with only 10-minute data tags, and announced that they continue their work on this issue. For years to come, the company looks set to ensure that its automatic captioning feature will be virtually error free.