Runway claims with Project Luxo that AI video has crossed the uncanny valley.

Runway claims its Project Luxo video AI crossed the uncanny valley, proving synthetic films can engage viewers without distracting artifacts.

Runway has published Project Luxo, a memo in which the company states that AI-generated video has crossed the uncanny valley, that threshold where a synthetic image becomes realistic enough for its flaws to become disturbing. The argument: the artifacts that previously betrayed these productions—fixed gazes, unstable faces, approximate lip-sync—are now rare enough for the viewer to remain immersed in the story rather than looking for errors.

To support its thesis, Runway submitted three entirely generated short films and a fictional advertisement to a panel from the audiovisual sector, including producers, actors, studios, press, and union representatives. The feedback focused on emotional resonance, engagement, character attachment, and narrative strength. According to the company, the verdict was unanimous: the films hold up.

Each short was created by a single person, in a matter of hours to a few weeks. The Rogue, by Felipe Orozco, took less than a month. Runway also invokes Luxo Jr., the Pixar short presented at SIGGRAPH in 1986, which shifted the industry's perspective on computer animation, and speaks of a 'Luxo Moment' for AI.

Finally, the company cites its advertisement for a fictional watch brand, credited with over 100 million views on Instagram in 48 hours and often reshared without mentioning AI. Runway acknowledges, however, that several questions remain open, from control and consistency to rights, consent, and transparency.