Judge refuses to dismiss YouTubers' DMCA suit against ByteDance over alleged YouTube scraping for AI training

What happened

On 2 October 2026, Judge Jacqueline Scott Corley of the U.S. District Court for the Northern District of California denied ByteDance Inc.'s motion to dismiss in Ted Entertainment, Inc. et al. v. ByteDance Inc., No. 3:25-cv-10933. The court held that Ted Entertainment and video creator Matt Fisher had plausibly alleged that ByteDance circumvented YouTube's technological measures to obtain videos allegedly used to train its generative AI systems, so their claim under DMCA Section 1201 may proceed. (This updates an earlier item that recorded only that the court's order had been filed, with the outcome unknown.)

Background

The creators allege ByteDance scraped large volumes of YouTube videos, bypassing protections including a rolling-cipher signature system, IP blocking and rate limiting, session-bound streaming URLs, CAPTCHA challenges, and proof-of-origin tokens, to build training data for AI products including MagicVideo and Seedance. The court rejected ByteDance's arguments that YouTube's measures regulate downloading rather than access, that individual copyright owners lack standing under Section 1203, and that the plaintiffs had not plausibly alleged actual circumvention.

Why it may matter to you

This is a pleading-stage ruling, not a finding that ByteDance actually scraped your videos. But it establishes that, at least at this stage, courts may treat YouTube's anti-automated-extraction measures as DMCA access controls — a potentially significant avenue for creators whose videos were allegedly collected for AI training datasets.

What is not yet known

The factual question of whether ByteDance actually circumvented YouTube's measures, and what content was obtained and used, will be litigated now that the case proceeds beyond the motion-to-dismiss stage. The ruling does not decide whether training AI systems on the videos constitutes copyright infringement.