🖼️ NEW BLOG: Final part of our #DisAIAMPLIFIED series! How can images boost fact-checking? 📷 Visuals = key evidence 🤝 Text + multimodal + LLMs 🔎 OCR when text hides in images 👉 kinit.sk/multimodal-f... #FactChecking #Misinformation #TrustworthyAI #NLP

💡 @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

⏳ Almost time for #RecSys2025! ✨ Our CEO Mária Bieliková & lead researcher Michal Kompan are co-chairs, and PhD student Santiago de Leon Martinez will run a workshop. 🙏 Big thanks to the organizers for a great program. See you in Prague! 🎉 👉 recsys.acm.org/recsys25/reg...

Join us in two weeks in Prague at CEDMO Café. Our Lead and Researcher Jakub Šimko is moderating "Fact-checking and AI Tools under Pressure" panel. 🗓️ See you on 23rd September 2025. 👇 Register via link in comments 👇

📖 Great resource from @ebu.ch: a handbook on C2PA (Coalition for Content Provenance and Authentication) and its role in building trust and countering disinformation. 👉 Also, don't forget to check possible interlinking with @vera-ai.bsky.social tools 🔧

🤖📝 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

Fact-checking isn’t just about text 📝 Visual data can make a big difference — even across languages. 🌍 Our #DisAI paper “Multimodal and Multilingual Fact-Checked Article Retrieval” is out now. Read here 👉 dl.acm.org/doi/abs/10.1... #ResearchPaper #MultimodalAI #FactChecking

What is AI ethics really about? Many imagine that AI ethics means teaching artificial intelligence to “distinguish right from wrong.” Want to learn how it really is? More in the thread 🧵👇 #sensaj #sensaiproject #aiethics #ethicsinai

AI was supposed to be fair… instead, it did something else 🤷 🆕 In our newest paper we look at the political views of LLMs, how these biases generalize, and how models can be steered. 👉 Read the full paper here: arxiv.org/abs/2507.22623 #DisAI #DisAIproject

Confused about what is the EU AI Act really about? Join our interactive workshop on September 11th. We'll look at how to unpack the rules, explore high-risk AI, and get practical tips for compliance without the legal jargon. 👉 More info: kinit.sk/everything-y...

✨ Remember Simon Ostermann, Daniil Gurgurov and Ivan Vykopal presenting their paper at #ACL2025? Now that the conference has concluded, their #DisAI paper is published in the ACL Anthology 📚 👉 Find it here: aclanthology.org/2025.acl-srw...

💡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

Our researcher Jakub Šimko joined this year’s Children’s University at @comeniusuni.bsky.social 🏫✨ He introduced kids to the importance of researching social media recommender systems 🤳 Big thanks to all curious minds who joined! 🙌 #AIAuditology

At #ACL2025, Jaroslav Kopčan showed that dictionary-based methods can rival LLMs. Especially for morphologically rich languages like Slovak, they can be faster, cheaper & more reliable than today’s hyped models 📄Paper 👉 lnkd.in/gANkikWS

Last Friday, we hosted Aashish Anantha Ramakrishnan 🚀, PhD @ Penn State & member of Prof. Dongwon Lee’s PIKE group. Fresh from #ACL2025, he shared how to make multi-modal LLMs more trustworthy — even in contexts full of metaphors, sarcasm & hidden intent.

Looking back at five days of cutting-edge AI & ML trends at AIDA Summer School, topped off with our Tech Lead Martin Tamajka leading the final workshop, where he, among other things, gave participants an overview of the AI on Demand platform within the #AI4Europe project.

👉 For our slovak followers: Sme kvôli umelej inteligencii lenivejší? Naša špecialista na etiku - Katarína Marcinčinová hovorí v najnovšej epizóde podcastu LIFE HACK o digitálnej lenivosti aj o tom, prečo je stále dôležitý človek. 🎧Celý rozhovor tu: bit.ly/3J4oQOt

🧐 How can NLP help fact-checkers make their work easier? Our researcher Marián Šimko shares insights from the #DisAI project in the recent #CORDIScovery Podcast episode. 🎙️ Listen to the full episode: bit.ly/cordiscovery... #disinformation #podcast #cordis

📈 Our Policy Specialist, Filip Veľký, spent time at #CERTH in Greece as part of the #lorAI project! The visit confirmed one thing: research, commercialisation and policy engagement aren’t extras — they’re essential pillars of a credible institute.

🚀 NEW BLOG: How LLM models are improving fact-checking? 💡 While embedding models can find similar fact-checks, generative LLMs can truly understand and select the most relevant ones. We tested 5 different prompting strategies. 🔗 Read more in the blog: kont.ly/e7a02280

🌍Introducing EUDHIT - initiative that promotes responsible innovation, helping start-ups, businesses, and other organisations develop IT systems that respect human values, fundamental rights, and individual autonomy. 🔗Learn more: kont.ly/85d3199d #EUDHIT

#ACL2025 recap At recent ACL 2025 Conference, our researchers presented paper in which they looked on how easy it is to generate personalized disinformation using LLMs. Read in more: bit.ly/misusing-LLMs ℹ️ The work was partially supported by #veraAI, #GEPERO and #AIAuditology projects

📑 How can organisations keep up with the flood of ever-changing legal texts? We teamed up with Seesame under the Hopero project to explore how AI can help monitor and analyse legal documents 💻 👉 Read full story here: kinit.sk/seesame-auto...

Our researcher Róbert Móro was one of the panelists during the June edition of Night of Ideas at the French Institute in Slovakia🌙🌃 Róbert talked more about our work in the DisAI Project and @cedmohub.bsky.social project to tackle disinformation🛡️ #disai #cedmo #disinformation

Measuring the behavior of social media recommender systems is at the center of our #aiauditology project at #KInIT. Jakub Šimko, our Researcher, recently presented our position paper at @ewafworkshop.bsky.social. Learn more in Jakub´s LinkedIn post: bit.ly/4obTXrB Read the paper: lnkd.in/eKiUVMjW

🎉 Celebrate GAAD 2025 with us! Join our free webinar on May 15 to explore how AI can support or challenge digital accessibility. 🕐 13:00–14:30 CEST | 💻 Zoom 🔗 Register: bit.ly/4iP0vIy 👀 Speaker details coming soon! #GAAD2025 #BeAIware #AIForAll

veraAI assets are now available on the AI-on-Demand platform. They were presented at the first ever webinar of the AI-on-Demand project.

Project partners from CERTH-ITI, @kinitsk.bsky.social & Fraunhofer IDMT will present veraAI and provide details about what's available via the 'AI on Demand platform' in a webinar on 3 April 2025, 3-5 CEST. “AI Meets Media: Tackling Disinformation with Cutting-Edge Innovation” Register here.

🚨 Exciting milestone alert! We're entering the third and final year of the @vera-ai.bsky.social To kick off this last phase, our researchers joined the entire team in Naples. It was a great opportunity to reflect on our progress, exchange ideas, and plan the next steps for the crucial final year 🚀.

Here's no. 4 in our @ebu.bsky.social webinar series. This one features @kinitsk.bsky.social researchers Ivan Srba, Aneta Žugecová and Domnik Macko presenting capabilities of #LargeLanguageModels (LLMs) in the context of tackling online #disinformation. www.veraai.eu/posts/ebu-we...

💡 As the use of AI-generated text becomes increasingly widespread, the need for reliable detection and benchmarking tools has never been more critical. The IMGTB framework from @kinitsk.bsky.social and FI at Masaryk University is designed to tackle this challenge. 👉 bit.ly/4gZ83Il

On 20 January @kinitsk.bsky.social's Ivan Srba, Aneta Žugecová & Domnik Macko talk about investigation of the capabilities of Large Language Models (LLMs) in the context of tackling online #disinformation. Join for this vera-ai webinar, organized by @ebu.bsky.social. tech.ebu.ch/events/2025/...