What it is
Automatic repair of damaged photos: remove scratches and noise, sharpen faces, colorize black-and-white into believable color. The AI "understands" the era, the clothing and skin tones, and fills in natural colors.
Where it came from
Colorizing B/W has been around for a while (colorizer neural nets), but in 2024-25 prompt editors and dedicated services (Fotor, VanceAI, Picsart, plus "restoration" in Nano Banana) pushed the result to "wow" in a single click — and the trend went mainstream as a family, nostalgic thing.
Why it took off
- Strong emotion. A revived photo of your grandmother or grandfather is personal and viral at the same time.
- One click. No Photoshop and no retouching skills required.
- Social media. The "before → after" format is a perfect fit for Reels/Shorts and racks up reshares.
How to use it right now
- Free: online restorers, or Nano Banana/GPT Image with a phrase like
restore this old photo: remove scratches and noise, increase sharpness, colorize naturally for the period. - For quality and large sizes — upscale (Magnific, Krea AI, Topaz).
- Want to "animate" it (a blink, a smile)? That's the neighboring trend (see "Photo animation"). What you get: a clean, high-resolution color shot out of a creased B/W original.
What to watch out for
- Colors are the AI's guess, not a fact: a historical uniform or dress can come out in the "wrong" color. For an archive or museum, label it as a reconstruction.
- The AI can "invent" facial details — compare against the original so you don't end up with a different person.