← All work

Independent

He's Back

Role
Product Designer & Creative Technologist
Dates
Nov 2025

The Problem

The world of AI video has come a long way in just a few years. Early results often showed simple motion and distorted “melting” environments, but today’s models can generate far more realistic movement, facial expressions, and even lip syncing. AI video technology continues to improve at a rapid pace. Despite these advancements, many commercial uses still feel artificial. Movements can appear rubbery, characters may shift inconsistently from frame to frame, and the visuals can resemble footage with an overly aggressive social media filter applied. However, when used thoughtfully and with the right workflow, AI video can produce surprisingly convincing results. I created this short demonstration beginning with a real photograph of Rowan Atkinson (for demonstration purposes only), which was enhanced using AI image models. Several variations of the image were created, including different poses such as the final “back to the camera” shot. These images served as the keyframes for the AI video generation process. In traditional animation, keyframes are the primary drawings that define the most important poses in an action. Applying this same concept to AI video allows for greater control over the generated motion. The various scenes were put together in After Effects, with music created in Suno to add the final dimension to the video.

The Solution

The Impact