Karol Bagiński’s Post

Researchers from the University of Edinburgh have partnered with experts at Historic England to develop HAZEL, a pilot generative AI chatbot designed to assist authors of Historic England's guidance literature. Funded by the Digital Strategy Innovation Board (Historic England), HAZEL aims to improve the clarity, accessibility, and inclusivity of written content by suggesting revisions during the writing and editing stages. Key features of HAZEL include: ✅ Text summarization and analysis capabilities ✅ Suggestions for enhancing and extending content ✅ Updating language for improved inclusivity and tone ✅ Ensuring compatibility with text-to-speech software and other accessibility tools The team, led by Dr. Chiara Bonacchi (PI, University of Edinburgh), will also explore the ethical implications of creating HAZEL and its potential impact on the organisational culture, voice, and discourse of Historic England and other heritage organisations. Want to learn more about this exciting project at the intersection of AI, accessibility, and cultural heritage? Read the full announcement here:🔗 https://ow.ly/TNQr50RuNzJ #CulturalHeritage #GenerativeAI #Accessibility #Inclusion #HistoricEngland #UniversityOfEdinburgh #Innovation #Research

Today marks the end of mesic.ai by Česká asociace umělé inteligence and I’m proud to say that I was there not only on its first and final days but also for many other events along the way. It’s been an incredible experience packed with insights and learning, bringing me closer to my goals in AI and business. Here's a recap of the events I attended: 1️⃣ HolkyzByznysu 2️⃣ Panel Discussion: Czech Companies vs. AI 3️⃣ Creating a RAG Tool with Python 4️⃣ Practical Insights from AI Implementation 5️⃣ AI and Evil 6️⃣ Introduction to the EU AI Act: What It Changes for You 7️⃣ AI Transformation in Public Administration 8️⃣ AI in Czech Gastronomy 9️⃣ Leveraging AI in Business 1️⃣0️⃣ AI Ethics 1️⃣1️⃣ Parental Leave nor AI is not an obstacle Through these 11 impactful events, I’ve gained so many insights about AI and its role in the Czech business landscape. I’m excited to apply what I’ve learned not only at work but also in everyday life. I also want to express my deep gratitude to Lukáš Benzl and his team for such an incredible project and the enthusiasm they poured into it. I hope you can take some well-deserved rest now! 😊 #AI #AIinBusiness #CzechAI #BusinessTransformation #MesicAI #PersonalGrowth

Interested in Retrieval Augmented Generation (RAG) but tired of all the buzz and noise around it? Here's a great whitepaper from Dr. Paul Yu-Chun Chang and Bernhard Pflugfelder that cuts through the noise to provide clear insights from both industry and research. It features three main sections: Section 1: RAG Industrialization - Landscape and Strategy Section 2: RAG Recipes for Real-World Challenges Section 3: A Deep Dive into RAG Evaluation & Metrics #ai #RAG #llm #haystack #langchain

I'm excited to share that I have started the Explorer-level Artificial Intelligence course as part of the Talento Tech program, organized by the MinTic in Colombia. This program is a fantastic opportunity to deepen my knowledge and skills in AI, and I'm looking forward to applying what I learn to drive innovation and data-driven decision-making. Stay tuned for updates on my journey and insights from this course! #ArtificialIntelligence #TalentoTech #LearningJourney #Innovation #DataScience

A new white paper from #AI experts at appliedAI Initiative GmbH on Retrieval-Augmented Generation Realized: Strategic and technological insights for #industrial applications. Retrieval-Augmented Generation (#RAG) is a smart way to make AI systems, like #chatbots, more accurate and reliable. A simple breakdown: ▪️“Retrieval“ Think of it like a super-efficient search engine. When you ask a question, it looks through a huge amount of information to find the most relevant facts. ▫️“Augmented“ This means the AI gets a boost from the information it finds. It doesn’t just rely on what it already knows but also uses new, relevant data. ▪️ „Generation“ The AI then uses both its own knowledge and the new information to create a clear and accurate answer. Why it’s useful: ▪️Better answers By using real data, the AI is less likely to make mistakes. ▫️More control The information can come from trusted sources, so you know it’s reliable. ▪️Efficient It uses resources wisely, making it cost-effective. In simple terms, #RAG makes AI systems better at finding and using the right information to give you the best possible answers.

Missed the opportunity to see Yuval Harari live in Antwerp yesterday because I was attending my AI Business Architect class at PXL-NeXT. As I near the end of Harari’s new book Nexus, I’m more convinced than ever that AI’s evolution must keep the human element at its core. It's crucial to prioritize keeping humans involved in AI development to ensure that we maintain humanity in an increasingly digital world. Every week in my AI classes, I'm astonished by AI’s immense potential—but also reminded of the significant ethical and societal responsibilities that come with it. Let’s continue to innovate while anchoring ourselves in empathy, humanity, and purpose. The future of AI is bright, but it must always have humanity at its heart. This is what I will keep focussing on with Humanize IT. On a side note, I highly recommend reading Harari's new book. Whether or not you agree with its conclusions, it’s certainly food for thought and a great wake-up call to always keep the human in the loop, certainly with AI. #AI #HumanInTheLoop #HumanizingIT #DigitalTransformation #AIEthics

Hot news 💥 The next appliedAI Whitepaper is out which highlights the best practices in RAG applications. Together with many partnners, this time we created a more technical whitepaper filled with insights into RAG challenges and best practices. It complements the Make/Buy Whitepaper for LLMs of the appliedAI Institute for Europe gGmbH (https://lnkd.in/dJMk4nQw) which looks at this topic from a more strategic perspective. You like to apply RAG? Come talk to us. appliedAI is your partner to stay competitive in the Age of AI. Many thanks to the team! Bernhard Pflugfelder Dr. Paul Yu-Chun Chang

What’s real and what’s AI-generated? This week, I had the incredible opportunity to close the XIV International Congress on the Legal Status of Smart Cities at the University of Málaga in partnership with Savonia University of Applied Sciences. Together, we explored the transformative yet sometimes unsettling impact of Generative AI on our world. My keynote, “Do You Believe What You See?”, was aimed at empowering students and professionals alike to question what’s real and what isn’t in the age of AI. We delved into how AI is reshaping our workplaces and classrooms, but more importantly, we tackled an urgent issue: the fine line between authenticity and AI-generated deception in digital images. In a world where scams and misinformation are on the rise, understanding AI's role in image generation is no longer optional—it’s essential. Visual authenticity matters, and our ability to discern real from fake could have serious implications for trust in our online interactions. Thank you to both Universidad de Málaga and Savonia University of Applied Sciences for inviting me to share this knowledge. I’m grateful to help raise awareness on this crucial ethical issue and for the chance to spark discussions that will shape the future of ethical AI use. #GenerativeAI #SmartCities #EthicsInAI #KeynoteSpeaker #UniversityofMalaga #SavoniaUniversity #ArtificialIntelligence #AIandEthics #VisualAuthenticity #real #fake

We are super excited to participate in the African Evaluation Association (AfrEA) conference this month. This year, Wilma Wessels-Ziervogel will take to the podium to address the burning question: ''Are we experiencing a revolution in how we evaluate due to the emergence of AI?'' Here are some of the critical conversations she'll be diving into: 1️⃣ AI's Challenge to Traditional Evaluation Expertise 2️⃣ Reevaluating the Evaluator's Role in the Age of AI 3️⃣ Enhancing Participatory Processes with AI Insights 4️⃣ Merging AI Technicalities with Human-Centric Evaluation Skills 5️⃣ AI as a Catalyst for Reflective and Learning Practices in Organisations In Wilma's extensive thinking around the intersection between AI and the future of M&E, one thing is evident: AI has the potential to greatly enhance evaluation outcomes. However, caution is necessary: AI cannot replace real-world human insight and ethical discernment. So, evaluators must continuously reimagine their relationship to AI and the evolving evaluation landscape. Let's reflect on our role as evaluators: How do we harmonise the technical capabilities of AI with the essential soft skills that it cannot viably replace? We’d love to hear your thoughts! #SouthernHemisphere #AIinEvaluation #HumanCenteredApproach #EvaluatorsInnovation The Rockefeller Foundation Mastercard Foundation CLEAR-AA United Nations Population Fund (UNFPA) United Nations Environment Programme Finance Initiative (UNEP FI) Ford Foundation

Today we had the pleasure to welcome a delegation from the Treasury Department of the Italian Ministry of Economics and Finance at the #JRC Ispra site. We shared our research and engaged in conversation on the adoption of AI and emerging technologies by the Public Sector. The workshop's exchanges touched: the need to maintain an anticipatory approach when studying AI and emerging technologies, as also embedded in the JRC's approach; the importance of analyzing case studies on AI adoption across Europe, with an attention to share success stories, yet learning from the less successful experiences; the value of experimenting, to create public value and starting from the concrete needs of public administrations; an overview of concrete example of generative AI applications. #AIPublicSector #DigitalGovernance #PublicAdministration