Creating Viral AI Videos — How to Produce the Colossal Hand Effect
In this article, I’ll show you how to easily create one of the trending AI video effects currently dominating social media — the colossal hand effect, where a giant hand places iconic landmarks into a scene. I’ll also share practical tips and creative ideas that I haven’t published anywhere else.
These AI videos are entertaining by nature, but anyone with a marketing mindset can quickly see their promotional power. AI video creation like this can drive traffic to your social platforms, increase engagement, and even position your product or brand deeply in viewers’ minds.
Always remember: artificial intelligence is a tool in the creator’s hands. Working with AI is mostly technical, and as I’ll demonstrate — anyone can produce impressive results. However, when it comes to creating cinematic marketing videos, business promos, or meaningful visual storytelling — that’s where your personal storytelling skills make the difference.
So here’s what we’re going to learn today:
@liorshubert i call it the hand of god.... #paris #eiffeltower #torreeiffel ♬ Eiffel Tower - Symbol of Love (Live) - JOAVIC
What do we see in this video?
It starts with an empty lot. Then, a colossal hand descends from the sky and places the Eiffel Tower into position.
The same technique could be used with the Statue of Liberty, the Azrieli Towers, or even a real estate project under construction — which already hints at one of the most powerful applications of this concept. AI videos for real estate marketing are incredibly impactful. The same approach can also be used to create an AI video for Airbnb properties, boutique hotels, and vacation rentals, allowing hosts to showcase their property and surroundings in a unique and cinematic way. While you wouldn’t build an entire marketing strategy around a single clip like this, it can add a strong creative edge to specific platforms
How to Create the “Divine Hand” AI Effect
Step 1 — Creating the Images
The first step is generating two images: a starting image and an ending image.
Note that I actually begin with the ending image — the Eiffel Tower already placed in its location.
You can either use a real photograph or generate a realistic image with AI. The Google Nano Banana model can recreate famous landmarks with impressive (though not perfect) accuracy.
I generated the Eiffel Tower image using Nano Banana. The main reason was to achieve a specific camera angle, but I could have also used a real photo and modified its perspective with AI.
This is the prompt I used to create the ending image:
Ultra-realistic aerial photograph of the Eiffel Tower and its real surrounding district in Paris, perfectly accurate urban layout, true-to-life buildings, streets, gardens and pathways, natural daylight photography, captured from a high isometric angle, axonometric-style camera with parallel lines and minimal perspective distortion, extremely sharp focus across entire frame, realistic materials and textures, natural colors, subtle tilt-shift feel, slight miniature effect, crisp environmental detail, documentary realism, no stylization
Next, I removed the tower.
My first attempt used a simple instruction (“remove the tower”), but the result wasn’t satisfactory because Nano Banana created a new artificial surface where the tower had been — which didn’t match my goal.
So I wrote a new prompt emphasizing what should replace the tower (sand surfaces in all four corners).
This is the prompt I used to create the starting image:
Remove the tower. Leave the concrete surface underneath. Only in the four corners where the base of the magnifier is, leave 4 small sand squares.
Step 2 — Creating the Video
The final step is uploading the images into the AI model.
Pay close attention:
The second image you created (the empty lot) must be defined as the starting frame.
After arranging both images in the correct order, I used the following video prompt:
A colossal human hand enters gently from above, holding the tower between its fingers. The hand moves with slow, controlled, natural motion and carefully places the element precisely in its designated position on the ground.At the moment of contact, create a soft and realistic puff of fine dust that rises lightly and disperses near the base. The impact should feel heavy yet clean. No fire, no sparks, and no thick smoke.The hand releases the object and smoothly retracts upward as the dust settles.Maintain perfect lighting consistency and background stability. Photorealistic cinematic animation in 4K.
Professional Tips for AI Video Creation
Anyone aiming to create professional AI videos cannot rely solely on copy-pasting prompts. Prompts are great for learning, but achieving precise results often requires adjustments tailored to your needs.
Beyond simply reading and understanding the prompt, here are several important creative controls:
Hand Design Control – My prompt didn’t strongly define the hand’s appearance, but you absolutely can. You might prefer an alien-looking hand, a hand wearing gloves, or even mechanical claws. Adapt the design to fit your concept and branding.
Environmental Control – In my video, the environment remains static. However, you might want natural movement like passing cars or shifting shadows. Achieving this may require modifying the video prompt, though timing issues can make it challenging.
A sometimes more practical solution is introducing subtle changes in either the starting or ending image — for example, repositioning vehicles without altering the overall composition.
Grip Mechanics – The Eiffel Tower works perfectly with this prompt, but other landmarks (such as Sagrada Família) may challenge the AI model due to their complex structure. If you encounter issues, refine the prompt to describe exactly how the object is being held.
Landing Dynamics – Notice that the prompt defines how the object is placed and touches the ground. You may want to refine this depending on the object and environment. In my example, I emphasized softness and realistic reflections to enhance cinematic quality.