UncovAI at Datafluence 3: Fighting Information Warfare
On the dark web, synthetic voices, fabricated videos, and AI-written narratives circulate freely — and they don't stay there. They migrate into mainstream feeds, newsrooms, and boardrooms. On April 3rd, some of the people thinking hardest about this problem gathered in Lyon. Our CEO was one of them.
A Room Built for Hard Conversations
Datafluence 3, organized by the Centre interarmées des actions sur l'environnement (CIAE) at ILERI in Lyon, brought together military officers, entrepreneurs, and students around a single thesis: information warfare is no longer a peripheral concern. It's a core strategic threat.
Florian Barbaro, PhD, joined the closing panel alongside General Bruno Courtois, Colonel Bertrand B., Julien Lopizzo, and David Hornus. The conversation covered doctrine, detection, and the gaps that still exist between what organizations face and what they can actually do about it.
Most organizations are still not equipped to detect and qualify manipulated information in real time. — Key takeaway, Datafluence 3 closing panel
That line isn't a criticism of any single institution. It's a structural observation — and it describes the problem UncovAI was built to address.
Five Things Worth Taking Seriously
Five clear threads ran through the panel discussion. None of them are abstract.
Disinformation as asymmetric warfare
It's cheap to produce, expensive to counter, and increasingly difficult to attribute. That asymmetry is deliberate.
Data collection is a critical vulnerability
Reliable sourcing is the first line of defense. Corrupt the inputs, and every downstream decision is compromised.
Old doctrine, new relevance
Strategic frameworks developed before the internet are proving surprisingly applicable to modern influence operations. The tactics change; the logic doesn't.
Geopolitics and geoeconomics together
Influence strategies don't operate in purely political space. Economic interests shape narratives — and vice versa. Separating the two misses the picture.
The fifth point — civil-military collaboration — deserves its own paragraph. The consensus in the room was that no single actor, public or private, has the full picture. Effective responses to information threats require shared infrastructure, shared intelligence, and trust that takes time to build. Events like Datafluence exist precisely to accelerate that process.
The Gap Nobody Has Fully Closed
The most pointed observation from the panel was also the most practical: detection at speed. Organizations can often identify manipulated content after the fact — once a narrative has already spread, a voice clone has been deployed in a scam call, or a deepfake has circulated through a network. The damage is already done.
The ability to understand and respond to information threats in real time is becoming a core strategic capability — not a nice-to-have.
This is where the technology question becomes urgent. AI-generated scam and deepfake detection has matured significantly in the past two years, but deployment — particularly in live environments — remains uneven. Most enterprises, government agencies, and media organizations are running on reactive workflows when the threat requires proactive ones.
For high-stakes contexts like video calls and live meetings, real-time deepfake detection is no longer a research-stage concept. It's operational. The question is whether organizations will adopt it before an incident forces their hand.
Why These Conversations Matter
Datafluence 3 wasn't a product showcase or a policy briefing. It was a working session between people who rarely share the same room — generals, founders, and students who will inherit whatever infrastructure gets built now. That mix is uncommon, and it's valuable precisely because it is.
Florian's participation reflects something UncovAI believes: the technical work of building detection tools is only part of the job. Understanding the strategic environment those tools operate in — the doctrine, the threat actors, the geopolitical pressures — makes the technology more useful, not less.
Thank you to the CIAE, ILERI, and every participant who made this edition substantive. The dialogue is the point.
Detection Is a Strategic Capability
If your organization is still relying on manual review to catch AI-generated content, you're already behind the threat curve. UncovAI gives teams the tools to detect synthetic media — audio, image, video, and text — before it causes damage.
Get Started Free →