💡 @kinitsk.bsky.social researchers created MultiSocial, a new dataset to test how well tools can spot #AI-generated text on social media. 🔍 Fine-tuned detectors adapt best, and platform & language choice affects performance. 📎 bit.ly/4nuZoRc #VIGILANTProject #DigitalSecurity #Disinformation

🌍 #VIGILANTProject returned to #GLOBSEC2025 Forum in Prague! Brendan Spillane (@ucddublin.bsky.social) & @gary-munnelly.bsky.social (@adaptcentre.bsky.social) shared how VIGILANT Project can support authorities facing crises, extremism & radicalisation online. 👉 Learn more: bit.ly/4nwNAh8

🤖📝 Spotting #AI text on social media? @kinitsk.bsky.social's MultiSocial dataset (470k posts, 22 languages, 5 platforms) tests how well tools detect AI vs human content. Results show fine-tuned models work best—even small ones! 💡 🔗 Check it out: bit.ly/3VbrWTE #Disinformation

🛡️ #Disinformation isn’t just misleading—it can incite real harm. To combat this, the C5 Interaction Model, developed by TNO under #VIGILANTProject, gives law enforcement a powerful framework to #analyse and #counter disinformation before it escalates. 📄 Read the full report here: bit.ly/3HFFW56

📘💡New from #VIGILANTProject & TNO: a practical guide to understand & counter #disinformation It covers: 🔎 How it spreads 🌀 The C5 Model (Context, Causes, Content, Consequences, Cycle) 🛠️ Tools for impact & response 👉 bit.ly/45WYJki #DigitalSecurity #EUResearch

How can #AI help police face growing #disinformation threats? 📰 ‪@irishtimes.com‬ featured the #VIGILANTProject: ✅ Spot emerging risks like violent attacks & terrorism ✅ Analyse text & images ✅ Expose harmful networks 📑 bit.ly/4lBFq5S #DigitalSecurity #TechForGood

💡From 4–6 June, IDSF25 in Vienna brought together 500+ experts to discuss Balancing Sovereignty and Solidarity in the Digital Age. Representing #VIGILANTProject, Róbert Móro 🎙 (@kinitsk.bsky.social‬) joined the panel on digital threats and information integrity. 📑 bit.ly/41Hvpga #DigitalSecurity

🤔🔍 Unsure if content is real or AI-made? @edmo-eu.bsky.social suggests: ✅Search fact-checking sites or trusted media ✅Spot odd hands, eyes, backgrounds in images ✅Hear pauses/robotic tones in audio ✅Watch lip-sync issues/blurry edges in video See the full tips: bit.ly/4fDDHM0

It’s starting to feel like online #disinformation might be around every corner. What can be done?💡 In this new CORDIS podcast by Publications Office of the EU, Prof. Owen Conlan (‪@tcddublin.bsky.social‬) discusses how #EU research is helping address these threats. 🎧 bit.ly/45cHBXw

📰 #Disinformation is in the headlines almost every week—but what fuels it beneath the surface? The C5 Interaction Model, developed by TNO for #VIGILANTProject, helps police understand and respond to harmful information manipulation. 🎙️ Neill Bo Finlayson shares more 👉 bit.ly/4fhq1Gk #Security

🚨 Fighting disinformation with insightful analysis The C5 Interaction Model, developed by TNO as part of the #VIGILANTProject, provides a strategic framework for European police to tackle #disinformation linked to criminal activity. 🔗 Learn more: bit.ly/44XRB6K

📢 Published in Discover Global Society (@springernature.com‬): the C5 Interaction Model—a multidisciplinary framework for understanding #disinformation. 📖 Neill Bo Finlayson (TNO) explains the thinking behind the model 👇 🔗 bit.ly/3IglFmy #VIGILANTProject #Security #DigitalResilience

Harmful narratives don’t just shape opinions - they can fuel division and escalate tensions. By analysing how these narratives emerge and spread, we can better equip law enforcement to respond effectively. #VIGILANTProject #Disinformation #Security #DigitalResilience

🤖 In Poland, artificial intelligence emerged as a tool for political image crafting As @demagog.org.pl, a member of ‪@cedmohub.bsky.social‬, points out, this highlights the growing need for media literacy in the face of synthetic content entering political discourse. 📎 bit.ly/409b9n3

📌 Last week, the #VIGILANTProject took part in the Security Research Event 2025 at EXPO XXI in Warsaw, joining leading voices in #EU civil #security and #innovation ✨ Thanks to all who stopped by our booth and engaged with our work! 📎 More on VIGILANT Project: bit.ly/4ludmBQ #SRE2025 #DigitalTrust

💬 “We need to insist again and again that the difference between right and wrong is not arbitrary. If you negate this difference, you also call into question the conditions of mutual understanding.“ Zlatko Valentic (University of Freiburg) 🔗 Read the full conversation: bit.ly/3HXKO5g

💡 Last week, Neill Bo Finlayson attended the EPIC (European Police Research Institutes Collaboration) 2025 conference in Stavern, Norway to present some of the latest outputs from TNO for the #VIGILANTProject.

⭐ @gary-munnelly.bsky.social & Brendan Spillane (@adaptcentre.bsky.social‬, @tcddublin.bsky.social & @ucddublin.bsky.social) joined #GLOBSECForum2025 to share insights from the #VIGILANTProject 🔗 Find out more about the VIGILANT Project: bit.ly/4k3Hov8

🔍 #Disinformation doesn’t just stay online—it can trigger real-world disruption 🛡️ New article published in Discover Global Society (@springernature.com‬), developed within the #VIGILANTProject introduces the C5 Model to better understand its full lifecycle. 👉 bit.ly/3ZvV5LV #Security #C5Model

🗞️ ‪@irishtimes.com‬ features the #VIGILANTProject and its #AI-powered tools helping law enforcement detect online threats. Dr. Brendan Spillane, our Principal Investigator, explains how it tracks #disinformation linked to violence, like attacks on asylum centres. Read more 👉 bit.ly/43ig6e1

❗Far-right parties across Europe are using AI-generated images to fuel anti-immigrant and conspiratorial messages—aiming to stir emotions and win votes. 📑 A new @edmo-eu.bsky.social investigation reveals how #AI is powering these digital campaigns. 📎 Read more: bit.ly/3FnoKzW

🔴 This article by @edmo-eu.bsky.social, originally published by @maldita.es.web.brid.gy breaks down what #misinformation circulated after the Iberian blackout on April 28 and why it matters. 📖 bit.ly/4mhzccL

✨ Spotted in @irishtimes.com: Gardaí could soon use a new social media early-warning system to flag potential crimes before they happen. 🛡️💪 We’re pleased to see #VIGILANTProject mentioned in this article on how #tech is transforming crime prevention: bit.ly/43ig6e1 #Disinformation #CrimePrevention

🧭 “We must provide concrete tools and technologies that allow authorities to effectively combat #disinformation without compromising democratic principles,” says Zlatko Valentic from the University of Freiburg. 🔗 Read the full interview bit.ly/3CNoQ23 #VigilantProject #EthicsInTech

✅ On 24 March, the VIGILANT Project took part in the Horizon Europe Cluster 3 Infoday at the @uab.cat 👥 The event brought together researchers, legal experts, university staff, and end-users to explore how research projects like #VIGILANTProject can help address security challenges 🔗 bit.ly/4jKArzU

Hello Bluesky! The new Horizon Europe FORSEE project is launching here! We’re excited to expand our community here and explore how success in #AI can be defined and used to strengthen Europe’s position in the AI landscape. Follow us and find out more at: forseeproject.eu!

In February, the #VIGILANTProject participated in a CoU event co-organised by ACCIÓ and @cdti-innovacion.bsky.social 🧩 The event brought together researchers, and innovators to explore new #technologies in areas such as climate intelligence, emissions detection, and biomedical innovation.

🕵️‍♂️ How are #cyberattacks caught when prevention fails? 🔐 Even a short internet outage can cause serious disruption—so early detection is key. 🔍 Patrik Goldschmidt (@kinitsk.bsky.social) breaks down how data helps spot intrusions when defenses fall short: bit.ly/4jMfnbx #CyberSecurity #DigitalSafety

🤖 From fake news during elections to climate #misinformation, #AI is already making it easier for both state and non-state actors to shape narratives. The Special Report from @euractiv.com's #AICODEproject dives into how AI is reshaping the #disinformation landscape. 📖 bit.ly/4cKnDXr

📢 An analysis published by @icct.bsky.social explores how actors like ISIS and al-Qaeda are experimenting with #AI tools for propaganda, recruitment, and #disinformation. 📎 Read the full article: bit.ly/42shFHj #DigitalSecurity #Counterterrorism #VIGILANTProject

🧬 @europol-eu.bsky.social's latest #SOCTA25 shows how organised crime is evolving—leveraging #tech, digital platforms & instability to grow. To counter this, #VIGILANTProject connects researchers, police & tools to track harmful content & its wider #security impact. 🔗 bit.ly/3RSAInK

🔐 The @ec.europa.eu has unveiled #ProtectEU — a new strategy to address growing internal security threats 🔗 bit.ly/3RdRxJO At #VIGILANTProject, we’re proud to support this effort by helping police forces identify and respond to harmful content that undermines trust, safety, and democracy.

🛡️ On March 27, at the @europarl.europa.eu’s Special Committee on the European Democracy Shield, @domhajdu.bsky.social, Director of @globsec.bsky.social's Centre for Democracy & Resilience, shared key findings from the VIGILANT Project on evolving online #disinformation threats. 📺 bit.ly/3YjfSS3

📲 A report by @techvsterrorism.bsky.social shows how far-right terrorist propaganda is evolving online. 🔍 The report maps trends, platform blind spots, and the emerging threat of Generative AI in spreading #disinformation. 👉 bit.ly/42mYsp5 #OnlineExtremism #TechAgainstTerrorism #DigitalSecurity

🗳️🚨 Fake news & #scams have made their way into Greenlandic election campaigns. Misleading Facebook ads mimic real news to trick people into investment scams. 📑 Read the whole article by TjekDet and @edmo-eu.bsky.social: bit.ly/42fkEBy #FakeNews #ScamAlert #Greenland #Elections

🧠💬 Bridging the #AI knowledge gap for older adults! @tudublin.bsky.social and @adaptcentre.bsky.social have launched a project aimed at helping older adults navigate the world of AI, and address concerns like #disinformation, #privacy and #security. 📎 Read more about the project: bit.ly/3Y674io

🤖🔎 #AI is reshaping our world for the better—but like any technology, it can also be misused. The latest UNICRI and UNCCT report examines how AI could lower barriers for extremism, making sophisticated attacks easier to carry out, and how to prevent it. 🔗 bit.ly/42iVH9i #AIforGood #ResponsibleTech

📊 The @munsecconf.bsky.social explored #AI’s role in the 2024 #elections. Many feared chaos, but AI’s impact was smaller than expected. Why? ✅ Better regulations ✅ Tech safeguards ✅ Voter awareness ✅ AI’s own limits. 📑 Read the whole analysis: bit.ly/4ivWPMz

🔎 A new report by UNICRI and UNCCT looks at how law enforcement and counter-terrorism agencies in South Asia and South-East Asia can use #AI to tackle online extremist threats. 💡 It breaks down how AI can help, and what’s needed to ensure its use is both effective and ethical. 📑 bit.ly/3DRbDGn

Can #AI really sway an #election? 🤔 💭 A Briefing Paper by Centre for Emerging Technology and Security, and The Alan Turing Institute explores how AI-driven #disinformation played out in the UK, EU, and French elections (June–July 2024). 📖 Read more: bit.ly/43IgkNj

🤖 Fine-tuning #LanguageModels often comes with unpredictable performance due to randomness in training. 🎲 A new paper from @kinitsk.bsky.social & @vut.cz introduces Delayed Ensemble with Noisy Interpolation (DENI)—a novel method that reduces instability. 👉 Read more: bit.ly/43pjBRq #AI

❓Can #AI fight #disinformation? A new report explores how AI aids fact-checking, text analysis & detecting #deepfakes —while tackling bias & ethics. AI isn’t perfect, but it’s a key tool to support #factcheckers & build public trust. 📒 Read more: bit.ly/43nXr20 @meddmohub.bsky.social

⚖️ The EU #AIAct ensures #AI works for people, not against them, requiring impact assessments for high-risk systems like law enforcement & credit scoring. 💡 @kinitsk.bsky.social offers a simple way to help businesses stay compliant. 📑 Learn more in their latest blog: bit.ly/41izuGS #ResponsibleAI

🗳️ A new report from the Adria Digital Media Observatory project reveals how #GenerativeAI was used to spread #disinformation during the 2024-25 Croatian presidential campaign. Fake AI-generated content circulated mainly on Facebook and TikTok, often gaining significant traction. 💭 bit.ly/3Xfm6BW

🔴 #Disinformation is the digital plague of our time, and #AI is taking it to a whole new level. A new study by the Konrad Adenauer Foundation examines the rise of AI-powered disinformation, exposing key actors and real-world cases in Europe and Africa. 📑 Read the whole study here bit.ly/433ySHt

💡 Large language models: A game-changer in the fight against online #disinformation Researchers from KInIT explore how AI is reshaping the battle against misinformation. 🔍 🎥 Watch the webinar and discover key insights from #veraai, #VIGILANTProject, and #aicodeproject: bit.ly/431TrnA

🛡️ As part of the #VIGILANTproject, researchers are exploring new ways to detect machine-generated content. With #AI making #disinformation harder to spot, European police forces need advanced tools to analyze and respond to emerging threats. 📑 More information about the project: bit.ly/4jKArzU

🛑 #Disinformation is on the rise across Europe, but so are the efforts to counter it. The VIGILANT project is working closely with European police authorities to develop innovative tools that help detect and analyze harmful content. 💬 Find out more: bit.ly/4jKArzU @adaptcentre.bsky.social

🔎 How can police stay ahead in the fight against #disinformation? The #VIGILANTproject unites leading European #researchers and police to develop cutting-edge tools for detecting harmful content and understanding its social impact. 👉 Read more: bit.ly/4jKArzU @ucddublin.bsky.social

KINIT has achieved a 95% accuracy in detecting machine-generated text across many languages. By fine-tuning LLMs & employing statistical detection methods, they've enhanced the identification of AI-generated content, contributing to the fight against disinformation and misuse of AI: bit.ly/4fJ2PiV