Netflix Unveils AI Video Editing Tool That Removes Objects, Recreates Scenes

Netflix is moving beyond streaming movies to changing how they are made. 

The company recently introduced a new AI framework called VOID, which stands for Video Object and Interaction Deletion. While we’ve seen magic erasers for photos before, Netflix’s tool is doing something much more sophisticated: it doesn’t just remove an object; it recalculates the entire scene to account for that object’s absence.

Most AI editing tools simply patch a hole left behind when you delete something. If you remove a car from a crash, those tools might leave a weird blur or floating debris. VOID is different because it understands causality and physics.

If you use VOID to remove a person jumping into a swimming pool, the AI doesn’t just make the person disappear; it “re-imagines” the water so there is no splash at all. According to The Register, the researchers described VOID in a preprint paper as “a video object removal framework designed to perform physically-plausible inpainting in these complex scenarios.”

In another test, when a car is removed from a head-on collision, the AI calculates a path for the remaining car as if the accident never happened. The smoke, fire, and debris are replaced with a clean, undisturbed road.

Built on a stack of AI technologies

VOID is not a standalone system but a combination of several advanced AI models and datasets.

According to The Decoder, the framework is built on Alibaba’s CogVideoX diffusion model and uses synthetic training data from Google’s Kubric and Adobe’s HUMOTO. Additional tools include Google’s Gemini 3 Pro for scene analysis and Meta’s SAM2 for object segmentation.

An optional second processing step uses optical flow techniques to correct distortions and improve visual consistency.

Competing tools and Netflix’s claim

There are already tools in this space, including Runway, ProPainter, DiffuEraser, and Generative Omnimatte. But Netflix researchers claim VOID performs better in complex scenarios.

In a small internal study of 25 participants, VOID was preferred 64.8% of the time, compared to 18.4% for Runway.

“Through extensive evaluations… we show that VOID excels at modeling complex dynamics which can follow on from object removal,” the researchers wrote.

Netflix has made VOID publicly available under the Apache 2.0 license via platforms such as Hugging Face and GitHub. That means developers, studios, and researchers can experiment with it freely, even for commercial use.

However, the tool is far from lightweight. Running it requires high-end hardware, such as GPUs with around 40GB of VRAM, putting it out of reach for most casual users.

Also read: The best AI photo editing tools in 2026 show how quickly AI cleanup features are moving from simple image fixes to more advanced visual editing workflows.

The post Netflix Unveils AI Video Editing Tool That Removes Objects, Recreates Scenes appeared first on eWEEK.