UNDERCOVERLABGenerative AI has driven the cost of faking down to almost nothing. Today, producing text, voice, images or video that looks real is within anyone's reach. When the fake scales faster than the genuine, the business question stops being how we make new things and becomes how we prove that what we show is real.
The problem isn't new, but now it's structural
For years we relied on indirect signals to know who was behind a piece of content. Style, context, continuity, reputation. All of that worked for one very concrete reason: faking was expensive.
AI has broken that logic. Impersonating a voice, cloning a style or generating a believable document no longer takes time or budget. What used to be a one-off attack is now a routine way to operate. Impersonation stops being the exception and becomes the default case you have to rule out.
Today the fake scales faster than the genuine.
Without authenticity, security arrives too late
Here's the shift that hits your operation directly. If you can't verify that an action comes from who it claims to, security stops preventing and starts chasing damage already done. You block after the fraud, not before.
Verification flips the order. It establishes trust before anything happens, not after. For a business, that's the difference between avoiding a problem and managing its fallout.
Verifying doesn't mean exposing data
There's a common confusion worth clearing up. Verifying the provenance of an action doesn't force you to show sensitive data or anyone's civil identity. What gets checked is something else: the coherence between an origin, an action and whoever attributes it, through cryptographic signals.
Put in business terms, you can prove something is genuine without turning yourself into a warehouse of personal data you then have to protect and justify. Privacy and verification are not at odds.
Verification as an infrastructure layer
At its core, the problem is an economic asymmetry. The fake is cheap, fast and unlimited. The genuine is expensive, slow and finite. When two forces this unequal share the same space with nothing to separate them, the system tilts towards the noise. Whatever is cheaper to produce wins.
Verification is the layer that separates the two worlds. It isn't control or surveillance: it's system maintenance, like inspecting the structure of a building so it holds up. And it can be done in different ways depending on who you are:
- Platforms: minimal content traceability to tell real users, legitimate automation and malicious actors apart.
- Creators: proving authorship and protecting a style anyone can replicate in seconds.
- Organisations and products: adding a provenance layer to every critical interaction, where a mistake costs money or reputation.
And an important point for anyone worried about losing control: this doesn't mean centralising anything. The layer can be distributed, interoperable and based on open standards.
Reputation without traceability is an asset you haven't protected. When impersonation costs nothing, it stops being worth anything on its own.
The objections, out loud
It's worth putting the arguments against on the table, because each has a grain of truth and an answer.
They say it adds friction. True, if it's designed badly. But the alternative, an environment where everything is suspect by default, costs far more in time and lost trust.
They say it centralises power. It would, if you built it with a single gatekeeper. That's exactly why distributed models and open governance exist: to avoid it.
They say reputation is enough. That was true when impersonation took effort. Today it happens at scale and at zero cost, and reputation, with nothing to trace it, doesn't hold.
The parallel with HTTPS
We've been here before. There was a time when encrypting web traffic was an option reserved for banking pages. Today, operating without HTTPS is unthinkable: the browser flags the site as insecure and you lose customers before they even reach the home page.
Verification will follow the same path. Today it looks like an add-on; tomorrow it will be a requirement to operate. The difference will come down to who gets ahead of it and who only acts once there's no alternative left.
And here's the misunderstanding worth undoing. Verification doesn't limit creativity or freedom. It holds up the environment where both can exist without drowning in copies. The sooner it's built in, the more room there is to design it well instead of bolting it on in a rush.
At Undercoverlab we build product with verification and blockchain because we believe this layer will be part of the foundation, not an extra. It isn't a trendy bet: it's anticipating where the structure of the digital world is heading.
