We’re Hitting Like on Millions of Fake Photos. We Either Stop or the Damage Will Be Irreversible
The AI wedding photos of Zendaya and Tom Holland passed eleven million likes on Instagram. Posted on March 4 by a random creator. Completely generated by Artificial Intelligence. Eleven million people who hit like on an event that never existed. But who notices when the light is perfect and the setting is romantic?
January 2025, California wildfires. Someone generated AI images of the Hollywood Sign engulfed in flames and put them into circulation on X and Instagram. They spread within minutes. Authorities had to publicly announce that the landmark was untouched. Not to correct a news story, to stop the panic.
At the 2024 Met Gala, AI photos of Katy Perry were so convincing that her own mother thought she was there that evening.
We’re at this point now. No hacking required, no infiltrating anything. A few seconds and a free tool are enough to put a convincing photo into circulation: a crime that didn’t happen, a politician saying things they never said, a person in trouble for something they never did.
And the platforms? Meta cut its fact-checking teams while the problem was exploding. X deliberately lowered filters on Grok for generating images of public figures. OpenAI in 2025 updated GPT-4o’s policies to allow the creation of images of politicians and public figures on simple request. The people running these tools know exactly what’s happening.
We got used to not believing words. Now we have to learn not to believe images. Nobody prepared us for this.
#ArtificialDecisions #MCC