Siren’s legal AI tool featured in Arte disinformation report
7 Jul 2025
|
Blog
In an age of information overload, one kind of falsehood carries especially high stakes: misleading claims about what’s legal and what’s not. Whether it’s a political figure misrepresenting international law or viral posts twisting legal facts, this kind of disinformation can inflame tensions, undermine institutions, and distort public debate.
That’s why, with French media organisation Les Surligneurs, we built Marqueur: an AI-powered legal analysis tool to help verify legal claims in public discourse. And we’re proud that the tool was recently featured in a special report by Arte, the Franco-German broadcaster, as part of its coverage on the promise and peril of AI in the media space.
Filmed in our Beirut office, the segment highlights how Siren Analytics is combining artificial intelligence and legal expertise with a commitment to public good.
“We have different profiles,” explains Jessica Chemali, who leads the Marqueur project. “Here we have the AI engineers. Then there are legal experts and political scientists.”
Together, they’re building tools to support institutions—and the people who hold them accountable.
Fact-checking political claims, in real time
In one part of the feature, the Arte team tests a real-world statement:
“An American senator recently said that the International Criminal Court is not qualified to issue an arrest warrant against Benjamin Netanyahu for war crimes,” Jessica notes.
Using Marqueur, they input the statement and watch the tool go to work. “We can paste this declaration to see if it complies with international humanitarian law,” she says.
Unlike basic fact-checking tools, Marqueur doesn’t just offer a yes/no verdict. It cross-references statements against legal texts from international humanitarian law sources, surfaces relevant citations, and explains the rationale—giving journalists, researchers, and the public a powerful way to challenge misinformation.
A tool for better governance
In a country like Lebanon, where political discourse is often heated and the media landscape is polarised, tools like Marqueur offer a way to cut through the noise and strengthen accountability.
“The shared dream,” Celine Moghrabi, a legal specialist at Siren, says with a smile, “is really to reach a point where, when a politician says anything foolish on TV, we’ll have a product like this, perhaps, that will come and display ‘N’importe quoi’(‘Rubbish’) next to it.”
It’s an ambitious goal—but one we’re committed to. Because in the fight against disinformation, it’s not enough to call things out. We need to build tools that show why they’re wrong, in ways people can trust.
Building AI for the public good
While the Arte report also warns of the dangers of AI-powered disinformation, from deepfakes to fabricated narratives, it ends on a note of urgency: the need to educate and empower people to critically engage with information, not just detect its flaws.
At Siren Analytics, that’s exactly what we aim to do: build technology for systems that serve, and for people who lead change. Marqueur is just one of the ways we’re using AI to promote access to information, transparency, and informed public discourse.
