Revolutionizing Cinema: Act-1 Enables AI Character Manipulation

Runway's Act-1 revolutionizes cinema by enabling you to control AI characters, offering a groundbreaking interactive experience.

The New AI Video Revolution

Leading AI video platform Runway has just unveiled a groundbreaking innovation that’s set to transform character consistency and film production as a whole.

Act-1: Modern Puppeteering

The feature, dubbed Act-1, brings a fresh twist to AI video generators. It’s akin to modern “puppeteering”, allowing the filming of a person acting out a role and then using AI to drastically alter their appearance. This development tackles one of AI’s biggest hurdles: consistency.

Act-1 will gradually become more available in the coming weeks, with Runway promising widespread access soon.

Bridging Humans and AI

AI video tools have significantly enhanced human motion capture, lip-syncing, and character development. Yet, there’s still a gap in eliminating the “obviously AI” effect. Runway’s new tool appears to be the solution.

Instead of letting AI dictate character movements or reactions, Act-1 allows for uploading a video alongside a control image to set the style. It then overlays this control image onto the filmed performance.

A Promising Future for AI Video

I believe the future of AI video lies in blending generative AI with reality, rather than relying solely on AI. The best films already use visual effects combined with live-action shots, and AI is simply an extension of this approach.

Runway’s Act-1 emphasizes human performance, using AI as a filter. It positions humans as puppeteers, akin to Andy Serkis’s portrayal of Gollum in “The Lord of the Rings”—without the need for motion capture suits or expensive cameras.

While I haven’t tried it personally, Runway has shared some compelling examples. It seems you just need to sit in front of a camera and move your head. Elements of this technology have been around for a while, like at Adobe, but not with generative AI elements.

However, Act-1 goes beyond anything we’ve seen so far. According to Runway, “With Act-1, gazes, micro-expressions, pacing, and interpretations are all faithfully reproduced in the final generated product.” It even transcends simple puppeteering by creating complex scenes with third-generation AI video technology and integrating them with human performance.