Independent research archive · London · since 2025
Authenticity Crisis
When seeing, hearing, and reading are no longer enough.
Unverified subject You cannot tell whether this face belongs to anyone. Neither can a bank, a court, or a newsroom.
The face
A face no longer reliably proves a person.
The voice
A voice no longer reliably proves an identity.
The record
A video no longer reliably proves an event on its own.
This is not merely misinformation or a passing panic. It is a structural shift in how individuals, institutions, and societies determine what is real, because neither people nor organisations can reliably distinguish synthetic from authentic content without independent verification.
Documented. Not predicted.
$25M Fifteen wire transfers, authorised after a video call with colleagues who did not exist. Multinational firm · Hong Kong · real-time deepfake Feb 2024 $243K One phone call, in a voice the executive had heard a hundred times before. Energy company · UK and Germany · voice clone 2019 $1.1B Estimated deepfake fraud losses in a single year, roughly three times the year before. Industry analysis of publicly reported cases 2025Lost while this page has been open, at the 2025 rate
That last figure is the documented $1.1 billion estimate for 2025 divided by the seconds in a year. It is an average, not a live feed, and it is here because the annual number is too large to feel. Every figure above is documented in the incident archive together with its sources. Fourteen further cases are recorded there.
What it is
The Authenticity Crisis is the societal condition in which AI-generated media becomes functionally indistinguishable from authentic human-origin content under ordinary conditions of perception, ending automatic trust in media, communication, and identity verification.
Where this now stands: The EU AI Act transparency obligations for synthetic content have applied since 2 August 2026, and the United States began enforcing the TAKE IT DOWN Act notice and removal requirements on 19 May 2026. Ongoing developments are tracked in Signal.
Why it matters
Societies rely on trusted signals of identity, evidence, and communication. When artificial media becomes indistinguishable from authentic human-origin media, those trust systems begin to break down. Verification becomes necessary in situations where it was previously unnecessary.
Historically, certain forms of evidence were accepted as inherently credible. A photograph confirmed that something existed. A voice confirmed identity. A video served as documentation. Those assumptions formed the operational foundation of journalism, law, governance, and everyday communication. That foundation is now weakening.
How this hub is organised
The Authenticity Crisis is a permanent, publicly accessible reference that defines the concept, documents real-world incidents, and tracks verification infrastructure and institutional response.
Browse by purpose
- Report The flagship analytical report introducing the Authenticity Inversion Model and the cross-sector impact of the shift.
- Definition The formal definition, why it matters, when it began, and what it is not.
- Incidents The evidence base: documented cases across fraud, impersonation, electoral interference, and the liar's dividend.
- Essays Reference analyses on trust collapse, identity uncertainty, visual evidence, and verification infrastructure.
- Library Curated research papers, standards, institutional reports, and policy frameworks for citation and follow-up.
- Signal Structural developments in capability, provenance infrastructure, and regulatory timelines.
- About Scope, methodology, verification standards, editorial stance.
17 documented incidents (2019–2025) · 5 essays · 30+ research sources · 8 signals