Daily 8/18: ByteDance signs a copyright pact with the MPA, YouTube changes how it counts views, Higgsfield raises $400M

ByteDance and the MPA agreed to strengthen copyright safeguards in Seedance and Seedream, YouTube starts counting a view the moment playback begins on August 24, and Higgsfield raised $400 million at a $5.4 billion valuation.

• [Reuters] ByteDance and Hollywood Will Build Copyright Safeguards Into Seedance and Seedream ↗

  ByteDance and the Motion Picture Association agreed to strengthen copyright safeguards in the Seedance video model and the Seedream image model. Disney and other Hollywood studios raised the issue in February, saying the models could generate copyrighted characters and the likenesses of well-known people without permission. ByteDance has added stronger IP protections to its latest models, and both sides will keep developing the safeguards together. Seedance and Seedream also feed large consumer services such as TikTok, CapCut, and Dreamina, so the pact reaches straight into how rights get handled between AI video generation and existing content IP.

• [TechCrunch] YouTube Will Count a View the Moment a Video Starts Playing ↗

  Starting August 24, YouTube counts one view the moment a regular video or a livestream begins playing. The old standard stays in creator analytics as a separate metric called Engaged views, and that is still the number behind ad revenue and YouTube Partner Program eligibility. Shorts, TikTok, and Instagram already count views this way, so YouTube now tracks public view counts and actual watch duration as two different metrics.

• [Reuters] Higgsfield Raises $400M at a $5.4 Billion Valuation ↗

  AI video platform Higgsfield raised $400 million in a Series B led by DST Capital. The $5.4 billion valuation is four times what it was six months ago. The company says it passed $700 million in annualized revenue and 30 million users worldwide as of August, and that its customer base now reaches entertainment and broadcast companies alongside advertising and marketing. Investors are putting the large money into the application layer that stitches several generative models into a working production workflow, rather than into the foundation models themselves.