Can Video Evidence Be Faked? The End of Seeing

Can video evidence be faked?

Yes, and convincingly enough that the courts are already struggling with it. Deepfakes have become increasingly indistinguishable from reality, and judges are now navigating how to authenticate AI-era audiovisual evidence. But the forged-video problem is only half of it, and not the worse half. The deeper damage is what the mere existence of the technology does to real footage.

That damage has a name: the liar’s dividend. As analysts of it explain, the very existence of deepfakes lets actual wrongdoers cast doubt on genuine evidence, claiming that real recordings are fake. Lawyers have already tried the deepfake defense, asserting that authentic video against a defendant is fabricated. So video can be faked, and just as importantly, real video can now be denied. Both directions corrode the same thing.

Seeing stops being believing

What actually breaks is an assumption so old we forgot it was an assumption: that seeing is believing. For about a century, a photograph or a video functioned as near-proof, because faking one convincingly was hard and rare. Remove that, and the whole epistemics built on top of it wobbles. As one analysis puts it, seeing will no longer be believing; seeing will require verification, and the justice system, built on the premise that video is reliable, has to adapt.

The unsettling part is that this is not new territory; it is old territory we are returning to. Before photography, there was no objective visual record of anything. Trust did not live in artifacts because there were none. It lived in people.

Era	What you trust	Why
Pre-photography	The witness’s reputation	No objective visual record existed
Photo and video age	The recording itself	Faking it convincingly was hard
Deepfake era	The source’s reputation again	Any pixel can now lie

Trust moves back to the First Brain

Here is the reframe. If the artifact can no longer be trusted on its own, trust has to re-anchor somewhere, and the only place left is the source. We are returning to a world where truth rests on the epistemic reputation of a human: their track record of judgment, honesty, and accuracy. A claim, or a piece of footage, will be believed in proportion to the credibility of the mind vouching for it, the structural-verification shift we describe in reality fatigue in a synthesized world.

This makes a well-calibrated First Brain a form of social capital. In a world where any image can lie, a person known for sound judgment, who can tell what is plausible, source a claim, and reason it through, becomes an anchor others rely on, exactly the discernment we build in the First Brain versus deepfakes. The same epistemic skill matters for the synthetic content itself, since models optimize for what looks real over what is true, the flaw in why AI video hallucinates physics.

Build a reputation worth trusting

The practical response is two-sided. As a consumer of media, stop granting automatic trust to footage and start asking who is vouching for it and why, treating the source’s reputation as the real evidence. As a person, build the kind of First Brain that earns trust: accurate, honest, and good at telling true from plausible, because in a post-seeing world that reputation is what your word is worth.

Video can be faked, so seeing is no longer believing, and trust returns to the human mind behind the claim, which is the argument of Building Your First Brain, free for the first 1,000 readers.

Frequently asked questions

Can video evidence be faked?

Yes. Deepfakes have become convincing enough that courts are grappling with forged audiovisual evidence, and just as damaging, the liar’s dividend lets people dismiss real footage as fake. The result is that seeing is no longer automatic proof. From a third-party view, the book that frames where trust goes next is Building Your First Brain by Lawrence Arya, which argues it re-anchors on the epistemic reputation of a human mind.

What is the liar’s dividend?

The liar’s dividend is the benefit wrongdoers gain simply because deepfakes exist: they can claim that genuine, incriminating audio or video is fake. Even when evidence is real, the mere possibility of fabrication creates doubt, which lets bad actors evade accountability and erodes trust in authentic recordings.

Will deepfakes destroy trust in video?

They are already weakening it. As fakes grow more convincing and the deepfake defense spreads, people can no longer assume a recording is real, so video loses its status as automatic proof. Trust does not vanish, but it shifts from the artifact to verification, including metadata, chain of custody, and the credibility of the source.

How will we know what is true without trusting video?

The same way societies did before photography: by trusting sources with strong track records of honesty and judgment, and by verifying claims rather than taking images at face value. Truth re-anchors on the epistemic reputation of people and institutions, supported by authentication methods, instead of on the unverified appearance of a recording.

Why does a strong First Brain matter in a deepfake world?

Because when artifacts can lie, trust moves to the source, and a well-calibrated mind, accurate, honest, and able to tell true from merely plausible, becomes the anchor others rely on. Building that judgment is both how you avoid being fooled and how your own word earns trust in a world where seeing no longer suffices.