Deep down we all know Google trained its image generation AI using Google Photos… but we just can’t prove it.

I swear every time I see a new AI image from Google, there’s this weird déjà vu moment.
The poses, the lighting, the random vacation-style smiling family… it’s like my old Google Photos library got reincarnated as a model weight.

Like let’s be real:
Google collected billions of ultra-high-quality images for over a decade, neatly labeled with dates, faces, locations, context, even automatic tags. We all happily fed it our pets, sunsets, food pics, wedding albums, baby photos, memes… and Google said: “Thanks, we’ll improve Search ☺️”.

Fast forward to 2024–2025 and suddenly Google’s image generation is skipping ahead like it has the entire world’s photo history as a training ground.

But of course, they’ll never admit it.
Terms & Conditions? “We do not use your photos for advertising.”
Cool, but AI training isn’t advertising.

Deep down we all know what happened…
but we just can’t prove it.

(Edit using Chatgpt)