Runway's latest research reveals that AI-generated videos are becoming nearly indistinguishable from real footage. In a study involving 1,043 participants, only 9.5% could accurately identify AI-generated videos from genuine ones. Participants viewed 20 videos total—10 real and 10 generated—yet overall detection accuracy was a mere 57.1%, suggesting that most people struggle to differentiate between the two. The ability to identify AI content varied by video type; human-related videos showed slightly better detection rates, while animal and architectural clips were often mistaken for real. The advancements in AI video generation, particularly with the launch of Gen-4.5, highlight a significant shift in video authenticity perceptions. With most people unable to discern AI-generated content, the implications for trust and verification are profound. Runway emphasizes the responsibility that comes with this capability, advocating for transparency and collaboration in creating new standards for synthetic media. They acknowledge that while current standards like C2PA for metadata certification exist, they are not foolproof. Runway's commitment focuses on three principles: being transparent about their models' capabilities, collaborating with industry partners on verification standards, and engaging with creators and policymakers to establish norms for synthetic media. The findings signal a need for ongoing dialogue about authenticity as AI continues to evolve and integrate into media production.