Digital propaganda was already a force. With generative AI, it has become child's play. An image can invent an event. A video can fabricate a witness. An audio recording can put words in anyone's mouth. Falsehood has never been so easy to produce, nor so convincing to consume.
And it is journalism — the last line of defense between reality and fabrication — that finds itself on the front line. Not as an accomplice, but as a victim. Because when anyone can generate credible visual evidence in 30 seconds, the entire chain of informational trust begins to crumble.
The democratization of manipulation
Until recently, producing a convincing fake required resources. Video editing skills. Equipment. Time. It was the domain of intelligence agencies, Hollywood studios, or state-sponsored disinformation operations.
Today, tools like Nano Banana and dozens of alternatives allow anyone to generate photorealistic images of events that never happened. Voice cloning platforms can reproduce a CEO's voice from a few seconds of public audio. Video generators create testimonials from people who do not exist.
On the ground, I see teams destabilized. What was once considered visual proof has become a mere suggestion. Reality is blurring, silently, massively. A communications director at a major corporation recently told me: "We can no longer distinguish a real leak from a fabrication. We're paralyzed."
And this is not a theoretical problem. In 2024, deepfake videos of political leaders circulated during elections in over 40 countries. Fake CEO interviews moved stock prices. Fabricated audio recordings triggered diplomatic crises. Every month brings new cases.
Journalism facing an existential crisis
The role of journalism is undergoing a radical transformation. For decades, a journalist's value lay in their ability to access information and report it. Being where things happen. Showing what is happening.
But when "showing" is no longer enough to prove, the profession must pivot. The journalist of tomorrow will not be the one who reports — but the one who certifies. Their value will no longer be access to information, but the ability to validate it. Moving from "here is what happened" to "here is proof that this is indeed what happened."
This pivot is existential. Because newsrooms were already weakened by the economic crisis of the media industry. Staffs have been cut. Investigation budgets slashed. And now they are being asked to add a layer of sophisticated technical verification — metadata analysis, deepfake detection, source authentication — with ever-shrinking resources.
The paradox is cruel: journalism has never been more necessary, and it has never been less equipped to meet the challenge.
The liar's dividend: when doubt benefits the manipulators
There is a side effect even more pernicious than deepfakes themselves: the liar's dividend. The mere existence of deepfakes allows anyone to deny the authenticity of real content. An embarrassing video leaks? "It's a deepfake." A compromising recording surfaces? "Generated by AI."
This mechanism is devastating for democracy. It is no longer about creating fake content to pass off as real. It is about making the real suspect — indistinguishable from the fake. Generalized doubt systematically benefits those who have something to hide.
I observe this effect in the corporate world as well. Organizations are beginning to question the authenticity of internal documents, client communications, even meeting recordings. Trust in digital content is eroding at every level.
The defenses being built
The counterattack is organizing, but remains fragmented.
Source authentication: the Content Authenticity Initiative (Adobe, Microsoft, BBC, among others) is developing digital provenance standards. The idea: every photo, video, or document carries an encrypted certification chain, from the sensor to publication. Promising, but adoption remains slow.
Algorithmic detection: forensic analysis tools detect artifacts invisible to the naked eye in generated images and videos. But like spam detection, it is a perpetual race between detectors and generators.
Information blockchain: some newsrooms are experimenting with blockchain-based traceability systems to certify the integrity of their publications. The New York Times and BBC have launched pilot projects.
Media literacy: the ultimate defense remains the public's critical thinking. But in an environment where the sophistication of fakes exceeds most people's analytical capacity, relying on education alone is insufficient.
What businesses must anticipate
This problem is not limited to media organizations. Every organization that communicates — which means all of them — is exposed.
Reputational risk: a deepfake of your CEO announcing mass layoffs or a scandal can go viral before you have time to react. Do you have a rapid response protocol? Most companies do not.
Sophisticated fraud: fake video calls from executives have already been used to authorize fraudulent wire transfers. The Arup case in Hong Kong — $25 million lost via a deepfake video of a CFO — is just the beginning.
Source verification: in a world where any document can be fabricated, your due diligence processes, supplier verification, and regulatory compliance must integrate an authentication layer that most organizations have yet to build.
Credibility as a strategic asset
In a world where anyone can fabricate their own reality, the only scarce power becomes credibility. For media organizations, for businesses, for individuals.
A reputation for reliability — built over years of rigor, transparency, and honest reporting — is once again becoming a first-tier strategic asset. The brands that invested in trust will be those that best withstand the wave of disinformation. The media outlets that maintained their editorial standards will see their value recognized.
AI accelerates everything. Amplifies everything. And above all, it democratizes manipulation. But it cannot manufacture trust. That is our advantage. That is what remains irreducibly human in this equation.
I help organizations face these new informational risks — from crisis communication strategy to authentication protocols. If you want to assess your exposure, let's talk.