# Opinion: Label the fakes, and the real things too

URL: https://ai-pulse.aglaganova.com/blog/opinion-label-the-fakes-and-the-real-things-too
Published: 2026-10-09
Author: Roman Aglaganov, Full-stack developer
Category: Opinion
Region: Global
Tags: deepfakes, regulation

> AI-generated images and video are now realistic enough to fool almost anyone. Detection alone will not save us; we need a reliable record of where content comes from.

![Typographic cover: Label the fakes, content credentials for everything](https://ai-pulse.aglaganova.com/api/media/file/labels-1.png)

For most of history, a photograph was evidence. Not perfect evidence, as Soviet censors airbrushing people out of pictures showed, but good enough for most purposes. That era is ending. AI can now produce photos, voices and video that are indistinguishable from the real thing to most people, most of the time.

## Detection is a losing race

The instinctive answer is to build AI that spots AI. It helps a little, but it is an arms race the fakers tend to win: every detector becomes training data for a better generator. Hidden watermarks in AI output are useful too, but they can often be removed by cropping, compressing or re-recording.

## Provenance instead

The more promising approach flips the question from “is this fake?” to “where did this come from?”. The Coalition for Content Provenance and Authenticity (C2PA), founded in 2021 by companies including Adobe, Microsoft, the BBC and Intel, created an open standard called Content Credentials: a tamper-evident record attached to a file, showing which camera or app created it and how it has been edited, including with AI.

Some cameras and phones can now sign photos at the moment of capture, and some platforms display the credentials. Governments are moving too: the EU AI Act requires deepfakes to be disclosed, and China’s rules require labels on AI-generated content.

## What should happen next

Labelling only fakes is not enough, because the absence of a label proves nothing. The goal should be the opposite: real photos and recordings that can prove they are real. Until then, the most important tool is the oldest one: a healthy pause before you believe, or share, something that seems designed to make you furious.

## Sources

- [C2PA: Coalition for Content Provenance and Authenticity](https://c2pa.org)
- [Content Credentials](https://contentcredentials.org)