YouTube uses ML to automatically generate video chapters
YouTube is working on a new feature that will help generate video chapters automatically. The company is currently testing the feature with some users. There is no word yet on the official rollout of the feature.
YouTube on its support page has confirmed that it will use machine learning to help generate video chapters. The tool is likely to help reduce stress for creators to manually add timestamps.
We will use machine learning to recognize text to automatically generate video chapters. We are testing this with a small group of videos, YouTube said.
In case you didn’t know, YouTube rolled out video chapters in May of this year. The new feature allows viewers to quickly jump to the specific part of the video they want to watch. The tool could be useful for longer video formats such as explanations, round tables, interviews, etc.
Obviously, the creators haven’t fully embraced the new functionality. This can mainly be because the video chapters depend on the creators to insert captions with time stamps. The description can take up to 10 seconds.
A machine learning-based chapters tool could make it easier for creators to implement in their videos. That said, it remains to be seen how accurate these automatically generated chapters will be. In case that doesn’t work, creators may be more reluctant to use it.
Separately, YouTube faces a close review of its measures to control fake news, conspiracy theories and misleading content on its platform. A group of US Senate Democrats urged the company to remove videos containing “false and misleading” information about the election.
“These videos seek to undermine our democracy and cast doubt on the legitimacy of President-elect Biden’s new administration,” wrote Senators Robert Menendez of New Jersey, Amy Klobuchar of Minnesota, Mazie Hirono of Hawaii and Gary Peters of Michigan. Additionally, because the current president has not made a commitment to a peaceful transition of power, disinformation and media content manipulated on your platform can fuel civil unrest.
Besides YouTube, Facebook and Twitter have also come under intense scrutiny for managing this content on their platforms.