Video with AI: Reels, ads and translation

Video is the most watched content and the most expensive to make. AI speeds up most of the work.

What AI does with video

1. Editing and subtitles. Upload a video and AI cuts the pauses and adds subtitles.
2. Image to video. Upload one picture and it becomes a 5–10 second living scene.
3. Text to video. Describe a scene and AI creates it.
4. Translation. Your Uzbek video speaks Russian or English, in your own voice.

Step by step: a product Reel

1. Script. Ask Claude or ChatGPT:

Write a script for a 15-second Instagram Reel about a pistachio cake.
4 scenes; for each: what's in frame, on-screen text, voice-over.

2. Images. Create an image per scene (as in the design lesson) or shoot them on your phone.

3. Animate. Upload the image to Kling AI or Runway and describe the motion: “camera slowly moves closer, steam rises from the tea”.

4. Assemble. Put everything into CapCut, add music and generate subtitles with “Auto captions”.

5. Voice. If needed, generate a voice-over with ElevenLabs (see the audio lesson).

Tips

• AI video still works best for short scenes. Build longer videos from short pieces.
• Faces and hands sometimes come out strange — replace those shots.
• Don’t use other people’s faces or voices without consent.

Exercise

Make a 15-second video about one of your products or services and show it to friends. Which part took the longest? Hand exactly that part to AI next time.