Institution-Backed Verification
Verify synthetic-era claims through accountable institutions, not vivid media
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 5
- Confidence
- 97%
Harari argues that easy synthetic video removes a shortcut people previously relied on: believing an event because it looked or sounded real. His historical analogy is print. Any sheet of paper can attribute invented words to a prime minister, so confidence comes not from paper and ink but from the accountable institution that published and checked the report. The same rule can be applied to synthetic media. First detach the claim from its vivid format, then trace its source, seek an institution with verification processes and reputational accountability, and look for corroboration. The institution is not automatically correct; its value lies in identifiable methods, responsibility, and correction. When provenance or verification is absent, confidence should remain low even if the artifact appears emotionally or technically convincing.
Origin
Extracted from The Diary of a CEO. Yuval Noah Harari uses the transition from trusting printed artifacts to trusting accountable publishers as an analogy for verifying synthetic video.
Core principles
- 01A convincing medium is not proof of a claim
- 02Cheap fabrication shifts trust from artifacts to provenance
- 03Accountable institutions can verify what technology alone cannot
- 04Trust depends on a record that can be inspected and corrected
- 05Institutional trust is necessary but should remain evidence-based
How to run it
- 1
Detach Claim From Medium
Write down what the artifact claims without treating realistic video, audio, or typography as supporting evidence. Identify the consequence if the claim is false.
Pro tip Describe the claim in one sentence before sharing it.
- 2
Trace the Provenance
Find the earliest available publisher, original context, and chain by which the artifact reached you. Treat reposts and clips as leads rather than confirmation.
Pro tip Prefer the claimed speaker's official channel or a full primary record when available.
Watch out An official-looking account can also be impersonated.
- 3
Find Accountable Verification
Look for a publisher or institution that states how it verified the material and can be held responsible for errors. Evaluate its relevant track record rather than its familiarity alone.
Pro tip Check whether named reporters, editors, records, or methods are visible.
- 4
Seek Corroboration
Compare independent credible reports and primary evidence. Investigate meaningful disagreements instead of counting repeated copies of one source.
Pro tip Confirm that apparent sources are not all citing the same unverified post.
- 5
Calibrate Confidence
State confidence in proportion to provenance, verification, and corroboration. Delay sharing when the evidence does not match the claim's stakes.
Watch out This framework reduces risk; it cannot guarantee truth.
In the wild
The host plays an AI-generated clip that reproduces his appearance and voice even though he did not record it. Harari argues that a similar political clip should no longer be trusted because it looks authentic. The viewer should instead seek accountable verification and the artifact's provenance.
→ Visual realism stops functioning as the decisive evidence.
Illustrative example: an audio clip appears to show a chief executive announcing layoffs. Before acting, an employee checks the company's official channels, looks for the full recording, and seeks corroboration from accountable reporting. No source can establish provenance, so the employee withholds the claim.
→ The employee avoids converting a persuasive artifact into an unsupported fact.
Common mistakes
Trusting Technical Realism
Harari says video loses its special evidentiary status once convincing fabrication becomes easy.
Treating Repetition as Corroboration
Many reposts may still trace back to one unverified artifact.
Blind Institutional Deference
The mechanism depends on accountable verification and correction, not on assuming any institution is always right.
Is it for you?
Best for
It is best for citizens, journalists, teams, and leaders handling consequential claims in a synthetic-media environment.
Not ideal for
It is not ideal for blindly trusting a familiar brand or treating institutional publication as infallible proof.
From the transcript
“you cannot trust the technology you trust the institution that verifies the information”
“when it becomes very easy to fake videos then we revert to the same principle”
“you can't trust videos you can only trust the institutions”
From the episode
Yuval Noah Harari: This Election Will Tear The Country Apart! AI Will Control You By 2034! The Dark Truth Behind Meta & X!