← All media signals
Trust signal

Signed at Capture

Cameras are starting to sign a photo as real the moment the shutter closes, while most people still judge fakes by eye rather than by a credential.

What is it?

Cameras and editing tools are starting to sign what they capture as real at creation. The signing layer is growing across premium hardware, but almost nobody verifies by credential yet — people still eyeball a photo instead. Whether the credential ever reaches ordinary phones is the open part.

Why now?

Synthetic images got cheap and convincing enough that eyeballing stopped working, so the industry moved proof upstream, to capture rather than after the fact. A credential only means something if it survives every hop — and ordinary messaging and email still strip it the instant a file passes through.

Who’s building it?

Camera makers and standards bodies signing at capture — mainstream distribution and stripping remain unsolved above them.

Camera makersC2PA standards bodyNewsroomsPlatforms · XMetadata-stripping tools
The full read
Why now, who is building it, where the market goes, and what stays unsettled.
Get weekly →
01 · The market
Who’s in it
Builders — camera and software vendors shipping cryptographic signing at capture.
Buyers — newsrooms and platforms adopting credential display to defend their own trust.
Gatekeepers — messaging apps and email clients, which still strip the credential in transit.
What’s at stake
The verification layer and actual verifying behaviour sit in two separate places — infrastructure exists where the public habit does not.
A stripped credential offers no protection at all.
Smartphone capture remains the chain’s largest unclosed gap.
What’s unsettled
Whether signing ever reaches the phone in every pocket, not just premium cameras.
Whether ordinary apps will stop stripping metadata once it actually matters.
Whether public habit shifts from eyeballing fakes to checking a credential instead.
02 · How the demand matureda progression, not a timeline
Eyeballing
People crowd into forums judging fakes visually, unaware a credential layer exists.
Capture signing
Premium cameras and editing tools begin signing at the point of creation.
Display adoption
Platforms and newsrooms start surfacing the credential, turning invisible infrastructure into a visible signal.
Universal reach
Signing has to reach ordinary phones and survive messaging apps to matter at scale.
03 · Where the market goesthe forks, not a prediction
→Universal or premium-only — whether signing spreads to every phone, or stays a feature of cameras few people actually shoot on.
→An adversarial community is already openly publishing recipes for stripping or defeating credentials outright.
→Messaging and email apps could adopt preservation, or keep stripping the layer that matters most.
→Verification could shift from visual judgement to a routine credential check.
Everything here turns on distribution: a credential nobody’s phone can produce and nobody’s inbox preserves is infrastructure without an audience yet.
04 · The Authenticity Premium deskadjacent markets, moving now
The Synthetic Stamp
A labelling standard for deepfakes gets drafted.
A standard is drafted.
Tech Policy Press
Proof of Hand
Courts still hold the line on human authorship.
Courts hold the human-author line.
CNBC
Radical Transparency
A watermark now spans billions of AI images.
SynthID has watermarked over 100 billion images and videos.
InfoQ
05 · The Wire
01
Sixteen signals of adoption.
Eyesift ↗
02
Signing reaches premium cameras.
Editors Weblog ↗
03
iPhone 18 Pro’s Reference mode signs sensor data at capture
Gizmodo ↗
04
Canon launches paid C2PA provenance system for news organisations
RedShark News ↗
The panel on this signal
They agreed on
–The signing infrastructure and the public’s actual verifying habit are being built in two disconnected places.
–Smartphone capture is the unclosed gap in the whole chain.
–A credential is only ever as strong as its single weakest hop in the chain.
The security researcher disagreed on
–Metadata-based authenticity signing can be stripped through common editing and re-upload workflows already in wide use; adversarial removal capability may continue to outpace adoption, keeping the credential a minority signal rather than a default expectation.
Know more about the method →
✦How this signal was built
01 · Panel
A Delphi panel of 4 AI analysts, each reasoning independently.
→
02 · Debate
They challenge and revise each other’s reads over two rounds.
→
03 · Evidence
Every claim is tied to a cited, linked source on The Wire.
→
04 · Synthesis
A moderator writes it into one read, disagreement and all.