News

First Forum for Information Integrity

France24: A Closer Look at the Wikipedia Sensitivity Meter

Overview of the public tools developed as part of the PROMPT project

Le Monde: 'Infox sur Seine 2026' and the Whole-of-Society Approach in the Information War

A Look Back at the Conference Co-Organized by opsci.ai

Propagations, the podcast that decodes the information war

ArchiMag: AI to detect disinformation narrative patterns

Jordan Ricker and Rémy Gerbet present the PROMPT project and opsci.ai's work with Wikimedia

Trustworthy AI in defense: our contribution to the PEC white paper

The Pôle d'excellence cyber third volume of experts' views on AI for defense

Sciences et Avenir: Three French initiatives for greater ethics and reliability

Clément Bénesse presents OpenLLM and the corpora of public, free-to-use data for training language models

Le Point: Disinformation, is AI the evil and the cure?

Justin Poncet talks about the pollution of information space by hostile powers

What value should be attributed to disinformation? Our contribution to the PEC white paper

The Pôle d'excellence cyber second volume of experts' views on the fight against disinformation

Opsci endorses OSI's definition of "open source AI"

An acceptable, realistic, and pragmatic balance given the current state of things...

Meet us in Riga from Oct. 7th to 11th for NATO disinfo & social media workshops and the #Disinfo2024 conference

Contact us to discuss our projects with the Ukrainian PR Army and the PROMPT consortium

Challenges: Social networks, the echo chamber of opinion

"Getting moved by a cause or lending your support on Instagram, blocking celebrities, it's all happening online. It's an activism of the click, from your smartphone or your computer."

France2030: Opsci and the OpenLLM-France consortium winners of the ‘Digital Commons for Generative AI’ tender

The OpenLLM consortium develops sovereign, open source, multimodal generative AI

The consortium led by Opsci will carry out the work of the European Narrative Observatory to combat disinformation

As part of our winning PROMPT project, we will be developing tools for detecting and analysing manipulative narratives in 6 countries

TF1: #AgriculteursEnColere

How the conspiracy movement has taken hold on social networks

Libération: The quest for general artificial intelligence

"Respecialising models: taking generalist ones and re-training them on specific, delimited tasks, to better document and canalise them"

Wired: a fine-tuned version of Stable Diffusion based on public domain stills of Mickey Mouse

"People should be paying closer attention to where AI tools get their training data"

Science&Vie : The rise of climate scepticism

"At a time when the French are increasingly confronted with the direct consequences of climate change, sociological studies show that climate sceptic ideas are gaining ground in France"

Opsci joins public incubator AllIAnce alongside HuggingFace, Mistral

Governement launches public-private alliance to create ‘sovereign’ generative AI

First Documented Code Notebooks in French for Fine-Tuning LLaMa-1 and LLaMa-2

A notebook tested on a challenging instruction set of 2,000 excerpts from 17th-century French novels

Public Sénat: The new terrain of conspiracy theorists

"There are more and more of them in France, joining the battalions of climate sceptics who refuse to accept scientific reality"

Calling on the regulator: Social networks can't be a black box for researchers

To extend the scope of the DSA and on platforms to respond to an immense democratic challenge

Our contribution to the European Digital Service Act (DSA)

Delegated Regulation on data access

De Facto: When anti-climate ideas raise the online temperature

"Two studies published in 2023 by the CNRS Institute of Complex Systems and the Fondation Jean-Jaurès, which observe the resurgence of these theories online, point to an unprecedented upsurge"

RTS: The songs of climate scepticism

"Three out of ten French people subscribe to a climato-complotistic statement about climate dictatorship"

L'ADN: The combination of conversational agents

"New models combining several conversational agents capable of interacting with each other are captivating GPT-bros. But are they really useful?"

France Culture: The breakthrough of the climate conspiracy mongers

"On social networks, climate scepticism is giving way to an even more virulent cousin: a form of pure climate complotism. Are we witnessing an anti-environmental backlash?"

"Climate dictatorship", "Climate pass", "Great Reset": Unraveling the impact of climate conspiracy theories on public opinion

Daily survey of the trending narratives and evolution of the leading forces

La Tribune: Online support for nuclear energy

"Already very positive in France, the image of the civil atom has improved considerably in recent months."

Exploring public opinion through artificial intelligence

Challenges and novel approaches for social data studies

Leveraging online public opinion to revitalize participatory democracy for climate action

A consortium of 13 European research and social innovation actors in response to the Horizon 2023 call for proposals

Twitter verification: noblesse oblige

The commoditization of the blue tick could help to empty the network of its substance

L'ADN: The political influence of neurchis

"Sometimes they talk about Tintin, sometimes they talk politics. To the point of influencing an election? Maybe..."

Modeling the neurchisphère

AI-powered insights into the world of french memes

Decoding climate narratives during election time

A hybrid approach to uncovering online public opinion trends

Green Wedge

Mapping climate dissent in Europe and «netnographic» approach to the communities reluctant to transition policies

100 000 climate change memes

Supervised semiotic analysis: memes as a counter-narrative in conflictual online spaces

Green Wedge Tracker

Monthly monitoring of the European climate conversation in eight languages