Crowdsourced Captioning

Crowdsourced captioning is an approach to creating video captions or subtitles by distributing the work across multiple contributors rather than relying on a single professional captionist. This method can leverage diverse workers with varying language skills, hearing abilities, and motivations — including hearing-impaired users who can identify errors in existing captions, second-language learners who benefit from the exercise, and native speakers who can correct complex errors. Crowdsourced captioning typically builds on automatic speech recognition (ASR) output, breaking the correction process into micro-tasks like marking errors, editing text, and approving changes. While less accurate than professional CART services, crowdsourced captioning offers a more scalable and cost-effective solution for the vast volume of uncaptioned video content online.