Creating realistic AI-generated videos could soon become much easier thanks to a new technology that allows people to control movement using nothing more than a computer mouse.
The system, called Time-to-Move (TTM), gives users precise control over how people, animals, and objects move in AI-generated videos without needing powerful computers or expensive training.
The technology was developed by researchers at the Technion–Israel Institute of Technology.
The team introduced TTM at the International Conference on Learning Representations (ICLR) 2026, one of the world’s leading conferences for artificial intelligence and deep learning.
Today, AI can generate impressive videos from text prompts or images, but controlling exactly how things move remains a major challenge.
For example, if someone wants a bird to fly in a particular direction or a person to walk along a specific path, existing AI video tools often struggle to follow those instructions accurately. Users usually have to accept whatever movement the AI creates.
TTM changes that. Instead of writing complex instructions or using advanced editing software, users simply move their mouse to draw the path they want an object or character to follow. The AI then creates a smooth, natural-looking video that closely matches those movements.
One of the biggest advantages of the new system is that it does not require retraining existing AI video models.
Many previous methods need developers to train AI systems using millions of video examples, which requires enormous computing power, time, and money. TTM works as a plug-in that can be added to existing AI video models, making it much easier and cheaper to use.
The researchers say the technology adds almost no extra computing cost. This means smaller companies, researchers, artists, and everyday users may be able to create high-quality AI videos without needing access to the massive computing resources available to technology giants such as Google or Meta.
At the heart of TTM is a new technique called dual-clock denoising. Although the name sounds technical, its purpose is simple. It helps the AI balance two important goals at the same time: following the user’s intended movement while keeping the animation smooth and realistic. The result is a video that looks natural instead of robotic or jerky.
In testing, TTM matched or even outperformed existing AI systems that rely on expensive training. The videos showed more accurate movement while maintaining a realistic appearance.
The new technology also offers creative features that many earlier systems cannot provide. Users can change the appearance of objects during a video or even add entirely new objects to a scene while keeping the movement believable.
Although the technology is still in the research stage, it represents an important step toward making AI video creation more intuitive and accessible. As tools like TTM continue to improve, creating professional-looking videos may become as simple as drawing a path with a mouse, opening new possibilities for filmmakers, educators, designers, content creators, and anyone with a creative idea.
