Voice Tweets Can Now Be Transcribed Into Automatic Text, Including Indonesian

JAKARTA - Twitter users now have the option to view automatic captions for voice tweets. The voice tweeting feature, which the company launched in June 2020 without subtitles, has sparked criticism from accessibility advocates and users who say it has excluded people who are deaf or hard of hearing.

Voice tweets, which are available on the Twitter iOS app, allow users to record audio of up to two minutes and 20 seconds, then send it as a tweet. Starting Thursday, July 22, users can enable captions by tapping the closed captioning icon in the upper-right corner of the voice tweet, Twitter said.

Twitter said automatic captions were generated using Microsoft Corp. technology, and users could not edit the text.

“As part of our ongoing work to make Twitter accessible to everyone, we are rolling out automated captions for Voice Tweets to iOS,” Gurpreet Kaur, head of global accessibility at Twitter, said in a statement Thursday, July 22.

"While this is still early days and we know it won't be perfect at first, this is one of many steps we are taking to expand and strengthen accessibility across our services," added Kaur.

A Twitter executive last year defended introducing voiceless tweets without text, saying in a tweet that adding a transcript would delay the launch of voice tweets until 2021 or 2022. The tweet was later removed, and the company said it would address the complaint.

The executive, Chief Design Officer Dantley Davis, also responded to the controversy, tweeting last June that he appreciates user feedback. “Obviously we have a lot of work ahead to make Twitter more inclusive for people with disabilities,” said Davis

When Auto captions are available for voice tweets in English, Japanese, Spanish, Portuguese, Turkish, Arabic, Hindi, French, Indonesian, Korean and Italian. As one of the countries with the most Twitter users, Indonesian has indeed become one of the priorities.

Twitter is the latest social media company to add automatic captions, though it's not universally available. TikTok, operated by ByteDance Ltd., also introduced a similar feature in April, and Facebook Inc.'s Instagram. launched auto captions to the IGTV video app in September.

While automated captions can make products usable by more people, they can result in errors. “Accents and background noise, for example, can cause incorrect transcription and change the intended meaning of sentences,” says Meryl Evans, a marketer of deaf digital tools.

"Audio presents a challenge that video doesn't have, as there are no visual cues to help users understand if they encounter confusing or incorrect text," says Evans.

Advocates say the most important step a company can take is to think about being inclusive early on in the design process.

“Embedding accessibility early in a design's lifecycle allows you to avoid the extensive effort required to improve features that are already in production,” says Kate Kalcevich, who is hearing impaired and works as head of services at Fable Tech Labs Inc., an accessibility testing platform.