This marks the first major AI video generator deliberately avoiding copyrighted content, making synthetic media more accessible while removing legal barriers. The credit-based system combined with public domain training creates a precedent for unrestricted, widespread access to video manipulation technology that could enable unprecedented levels of deception without traditional copyright protections limiting distribution.
AI Video Model 'Marey' Debuts Using Only Public Domain Films
📰 What Happened
Moonvalley, a Los Angeles startup, has released Marey, an AI video synthesis model trained solely on public domain films. The product launched publicly in July 2025 after a limited March debut, using a credit-based system. Unlike other video AI systems that rely on copyrighted material, Marey claims to be '3D-aware' while using only public domain footage. The system has attracted attention from film industry veterans like VFX artist Ed Ulbrich, addressing both ethical and legal concerns surrounding AI training data.
📖 Prophetic Significance
The convergence of Marey's public domain approach with existing AI capabilities accelerates three prophetic vectors simultaneously. First, the removal of copyright restrictions enables wider synthetic media adoption, advancing the Rev 13:14 'image that speaks' capability. Second, the credit-based system establishes a digital control framework aligning with Rev 13:17's economic controls. Third, the '3D-aware' technology combined with public domain training creates a new pathway for mass deception (Matt 24:24), as historical footage can now be seamlessly manipulated without legal constraints. This convergence removes key barriers to synthetic media adoption while establishing economic control mechanisms.