By skyforbes We are afraid that AI will generate perfect fake videos that can be used to scam people, destroy reputations or manipulate elections. Deepfakes are genuinely frightening. But that is only the first half of the nightmare.
The second half is worse. Once AI is powerful and common enough, not only does it create fake clips, but also hands liars a ready-made excuse that can cover anything.
A politician is filmed taking a bribe: “Deepfake, an AI did it.”
A celebrity is exposed in an old recording using racist language: “Must be AI, the technology is crazy now.”
A public official says something idiotic, then later claims: “That was an AI voice clone, not me.”
Even when a video is real, people will have a harder time proving it. Detection tools exist, but research shows systems that spot deepfakes usually lag behind systems that generate them. Experts have started calling this the “liar’s dividend”: once realistic fakes are possible, dishonest people can dismiss inconvenient evidence as fake and gain an extra layer of protection.
From 2025 to 2030, the main danger may not be killer robots. It may be a decade of radical doubt. AI will not “kill” us with nuclear bombs. It will quietly attack our ability to agree on what really happened.
The biggest problem is bigger than fake content.
If AI can fake anything, then nothing is real. If everything could be fake, many people will stop believing in anything. In the past, we read the news to learn about the world. In the future, some people may feel they have no knowledge and no reason to log on at all.
We will have to learn how to live in a world where sight and sound are no longer automatically trustworthy. Evidence from our eyes and ears will need cross-checks, context, and verification.
The Information Apocalypse has already started in a soft way. We cannot fix it with technology alone. New tools can help, but the deeper defence is something AI cannot manufacture: human judgment, conscience and a shared commitment to reality.