We were taking a family photo last weekend.
Third attempt, and my three year old was still blinking, so I kinda gave up and let the phone fix it.
One tap got everyone smiling and looking at the camera.
I was about to send it to the family group when I noticed my uncle's smile.
I had seen it a minute earlier, when he was laughing at something else.
The phone had taken his face from that moment and put it into this one.
The evening in that photo never happened, and that’s why I hesitated to send it ahead.
That is the part I want to talk about.
The feature is called Best Take.
When you press capture, your camera grabs a burst of frames, and Best Take lets you swap anyone's face for a better expression from a different frame.
So you get one image where everyone is smiling together, stitched from moments that never happened together.
Now, phones have been processing photos for years and none of it bothered me.
Night mode and HDR stack several frames to pull back the light that was in the room.
That is correction, and it shows you what your eyes saw.
But Google’s manufacturing something that never happened, its own marketing says that a customer they feature explains that when he hands his phone to a stranger now, "it doesn't matter if it's good or not", because Best Take can fix it later.
Whenever I bring this up in class, someone says the same thing.
"So what, people have been faking photos forever."
And that’s fair when you think about it.
If you go to a wedding, the photographer moves you two steps left, while the software they use (darkroom) lightened faces and probably removed a pimple when requested.
I agree with him.
The frame where you are mid-blink is not a truer record of your parents' anniversary than the one where you are laughing.
So if a photo exists to help you remember, Best Take is just a better remembering machine.
But nobody tricked me and that’s the whole problem.
Give anyone the version where everyone is beaming and the version where their father blinked, and watch which one survives.
Which means the honest photo was never something we valued.
The blink, the bad angle, the aunt looking away, we thought those were the flaws but that’s what kept the photos authentic and was the real “candid moment.”
We are now paying money to AI to have the authenticity removed.
For 150 years, the photograph has been doing a public job that nobody ever paid it for.
It was the cheapest evidence in the world.
If you have money, you have lawyers and contracts.
If you don't, you have a photo on a cheap phone, and that was enough to make a powerful person nervous.
A fake photo can be passed off as real, which is what everyone worries about.
The harder problem runs the other way, because once everybody knows a perfect fake is free, a real photo stops working as proof.
Lawyers already have a name for it, the liar's dividend.
So the powerful get a free defence, which is to look at a photo of the thing they did and say, probably AI.
And the woman with one photo of what was done to her loses the only witness she could afford.
In May, Google announced you will soon be able to ask Search or Chrome whether an image was made or edited by AI.
In that same announcement, Google said identifying authentic, unedited content "can be just as important as knowing when a file was made or edited using AI."
But if you can edit the photo, you can edit the AI label.
Half of them remembered the ride.
In detail, with feelings.
Your photo album today decides what your life was.
We just handed that job to a company that says its goal is a photo faithful to your memory.
So, one question: would you use AI to fix the picture if someone’s eye was shut or not?
Reply and let me know!
Nandini