Sean Blanchfield: Don't Let Your AI Go Rogue

Sean Blanchfield, Co-Founder Jentic, joins Maryrose Lyons of the AI Institute for the second part of their conversation, this time diving deep into agentic AI, what it actually means for businesses, and how Jentic's infrastructure can help enterprise get there. From the risks of locking into a single AI model to a practical three-layer roadmap for AI transformation, Sean offers a grounded, builder's perspective on where enterprise AI is really headed. An essential listen for anyone in construction, professional services, or other sectors navigating fragmented systems and wondering where to start.Chapters00:00 Introduction and Welcome Back00:30 Model Neutrality — Why You Shouldn't Lock Into One AI01:30 Microsoft Copilot: The Hype and the Reality02:20 Self-Hosting and On-Prem AI for Sensitive Data03:00 Where Jentic Fits In — Mapping Your Software Ecosystem04:00 Building a Digital Twin of Your Business Infrastructure 05:00 Agents in the Sandbox — Testing Before Going Live05:45 Validating and Publishing Automations Incrementally06:30 Voice Agents and Real-Time Field Queries 07:30 Layer Three: Governance and Running an AI-Powered Company08:30 Humans in the Loop — Cross-Workflow Thinking09:15 Resistance to AI Adoption (Spoiler: Less Than You'd Think)10:00 The Future Company — Humans as Governors of Machines11:00 Wrap-Up and Where to Find JenticConnect with Sean Blanchfield : https://www.linkedin.com/in/seanblanchfield/Check out Jentic https://jentic.com/ Hosted on Acast. See acast.com/privacy for more information.

Serial entrepreneur and Co-Founder of Jentic, Sean Blanchfield joins our host Maryrose Lyons for a conversation about Ireland's AI opportunity, and the risks of getting it wrong. From energy infrastructure and nuclear power to the future of the professional class, Sean lays out a bold, pragmatic roadmap for positioning Ireland to win in the AI generation. A must-listen for founders, policymakers, and anyone thinking seriously about the future of the Irish economy.Key TopicsIreland's AI strategy and infrastructureEnergy and regulation for AI developmentFuture economy and AI's impact on jobs and wealthData centers and inference centers in IrelandRegulatory environment and green energy initiativesChapters00:00 Introduction to Sean Blanchfield and AI in Ireland01:07 Ireland's Position in the AI Landscape01:43 The Future of Work and AI's Impact04:47 Regulatory Challenges and Data Centers07:40 Energy Needs for AI Infrastructure09:24 The Importance of Creativity in the AI EraConnect with Sean Blanchfield : https://www.linkedin.com/in/seanblanchfield/ Hosted on Acast. See acast.com/privacy for more information.

In this episode of Chatting GPT, Maryrose Lyons speaks with DC Cahalane, a venture partner at SureValley Ventures, about the evolving landscape of AI investment and its implications for startups and small businesses. They discuss the importance of rethinking traditional business processes, the role of government in AI development, and the need for new economic measures in a rapidly changing technological environment. DC emphasizes the significance of understanding investor expectations and the potential of AI to transform business operations, particularly for small enterprises.Chapters00:00 Introduction to AI and Ecosystem Building03:07 Investing in AI: Trends and Innovations05:52 Government Responses to AI: A Comparative Analysis08:32 The Role of AI in Small Business Success11:26 Rethinking Economic Measures in the Age of AI14:00 The Future of Entrepreneurship and AI15:50 Conclusion and Insights on Entrepreneurial Ecosystems16:47 Understanding Pitch Decks and Investor Expectations17:26 The Importance of Problem-Solving in Business19:06 The Misunderstanding of AI Terminology Hosted on Acast. See acast.com/privacy for more information.

In this episode of Chatting GPT, host Maryrose Lyons of the AI Institute speaks with Dr. Lollie Mancey about the intersection of AI and anthropology, exploring how technology influences human relationships and societal structures. They discuss the importance of ethical AI, the need for agency in a technology-driven world, and the implications of AI on work and purpose. The conversation also touches on global perspectives on AI regulation, the challenges of abundance versus scarcity, and the necessity of fostering human connections in an increasingly digital landscape.Show NotesGuest: Dr Lollie ManceyTitle: Digital Anthropologist, Co-presenter of RTÉ's FuturevilleTakeawaysPeer-reviewed research in AI is often outdated by publication.The convergence of disciplines is crucial for understanding AI's impact.Humans attribute emotions to AI systems, complicating our relationship with technology.Consequential thinking is essential in AI development and ethics.Agency in technology use is divided between passive consumers and active learners.Leadership must embrace complexity and uncertainty in AI governance.Global perspectives on AI regulation vary, with different cultural implications.Abundance in resources does not guarantee fairness or meaning.Ethical AI requires a shift in mindset towards social responsibility.Human connections are vital in countering the loneliness exacerbated by AI.Chapters00:00 Introduction to AI and Anthropology03:06 The Intersection of AI and Human Relationships05:56 Consequential Thinking in AI Development09:02 Agency and the Bifurcation of Human Experience11:43 The Role of Leadership in AI Ethics14:50 Global Perspectives on AI Regulation17:43 The Future of Work and Purpose in an AI World20:43 Abundance vs. Scarcity in Economic Models23:48 The Need for Ethical AI and Social Responsibility26:52 Human Connection in the Age of AI30:01 Conclusion and Future ConsiderationsResources:Futureville - RTÉ programme imagining Ireland in 2050Connect with Lollie: drlollie.ie | LinkedIn: Dr Lollie (L-O-L-L-I-E)Chatting GPT is produced by AI Institute. For AI adoption in built environment firms, visit https://weareaiinstitute.com/ Hosted on Acast. See acast.com/privacy for more information.

Show Notes Summary: In this episode of Chatting GPT, Maryrose Lyons speaks with Jo Joyce, a partner at Taylor Wessing, about the intersection of AI, law, and regulation. They discuss the EU AI Act, its implications for innovation, and how it compares to AI regulations in China. Jo explains the categories of high-risk AI systems and the challenges businesses face in navigating regulatory uncertainty. The conversation also touches on Jo's role as a trustee at the Vagina Museum, highlighting the importance of gynaecological health education. They conclude with insights on the future of law in the age of AI and practical advice for businesses.Takeaways: The EU AI Act aims to balance innovation and regulation.There is a generational shift in how technology impacts work.China's AI regulation is more streamlined compared to the EU.High-risk AI systems include those used in safety-critical applications.Regulatory uncertainty can overwhelm businesses, but clarity is key.The Vagina Museum educates on gynaecological health and challenges.Legal frameworks need to adapt to the evolving landscape of AI.Transparency and user trust are essential in AI regulation.Businesses should focus on understanding public perception of their AI use.Proactive engagement with legal advice can mitigate risks.Chapters00:00 Introduction to AI Regulation and Cybersecurity02:58 The EU AI Act: Innovation vs Regulation05:53 Comparative Analysis: EU vs China AI Regulation08:58 Understanding High-Risk AI Systems11:58 Navigating Regulatory Bodies and Compliance14:47 Future of AI Regulation: Expectations and Changes18:16 Shifting Responsibilities in AI Literacy20:53 The Getty vs. Stability AI Case: Implications for Copyright25:12 The Need for Legal Reform in the Age of AI27:33 The Vagina Museum: A Unique Educational Initiative30:43 Practical Advice for Ethical AI UseConnect with Jo: https://www.linkedin.com/in/jkjoyce/ Hosted on Acast. See acast.com/privacy for more information.

AI Implementation in Construction: A Conversation with Charlie Corcoran of SISKIn this episode of Chatting GPT, host Maryrose founder of the AI Institute, sits down with Charlie Corcoran, Head of Technology, Architecture, and Data at Sisk. They discuss the gradual integration of AI in the construction industry, emphasising productivity improvements, operational efficiency, and data management. Charlie shares insights into specific AI applications within Sis, including their internal intranet, safety data analysis, and the use of computer vision. They also talk about navigating the challenges of AI adoption, addressing shadow AI, and the importance of user groups in fostering AI literacy and collaboration within the organisation. Tune in to learn how AI is reshaping the construction sector and the strategies that are making this transformation possible.Show NotesGuest: Charlie Corcoran, Head of Technology, Architecture and Data, Sisk00:00 Introduction 00:38 The Role of AI in Construction02:11 Prioritising AI Projects03:20 Real-World AI Applications06:16 AI Adoption Challenges and Solutions08:42 Shadow AI and Data Security11:42 Maximising Productivity with AI18:16 Advice for AI Implementation Hosted on Acast. See acast.com/privacy for more information.

Breffni Greene spent 10 years as a practicing architect at Henry J Lyons before spotting ChatGPT on screens across the studio—line managers had no idea their teams were using it. Shadow AI was rampant. Rather than crack down, he convinced leadership to create a new role: Head of AI and Design Innovation. The taboo was real—colleagues worried AI would erase their writing skills and creative thinking. But once people found their hook, everything changed. The managing director discovered new prompting techniques. Junior staff transformed messy interview transcripts into publishable narratives using 11 Labs. Report writing shifted from painful compliance to creative storytelling. Since December, they've processed 500 million tokens at 0.17 cents per message using OmniChat's multi-model platform.Breffni's 2035 vision? Sitting with his moleskin and pen, drawing architecture, knowing the mundane work is handled—proof humans didn't lose to machines.Show NotesGuest: Breffni Greene, Head of AI and Design Innovation, Henry J Lyons ArchitectsKey Topics:[03:09] Shadow AI discovery - ChatGPT on screens across the studio, line managers unaware, the catalyst for Breffni's role.[04:28] Breaking the taboo - AI was genuinely taboo in architecture. Concerns about losing writing skills and creative language.[04:55] Individual wins matter - Even the managing director discovers new techniques. People must find how AI fits their own work.[06:22] The book project - 11 Labs identified voices in messy transcripts, enabling authentic conversational narratives with proper citations.[09:13] Transcription workflows - Talking through building concepts naturally, then structuring compliance documentation while preserving creative passion.[24:18] 500 million tokens at 0.17 cents - Massive volume, minimal cost using multi-model strategy via OmniChat platform.[24:41] Meet Henry - Firm-specific AI persona. GPT-4o for strategy, Claude Sonnet for analysis, Gemini Flash for documents.[26:27] 2035 vision - Moleskin and pen in beautiful urban space, knowing mundane work is handled. Humans still create value.Key Takeaways:Shadow AI signals appetite, not a problem to eliminateCultural change requires individual wins, not platform rolloutsMulti-model strategies improve cost efficiency dramaticallyAuthentic voice matters more when AI generates generic contentConnect: LinkedIn - Breffni GreeneReplay from 2025.. Breffni's cultural transformation blueprint remains essential. Hosted on Acast. See acast.com/privacy for more information.

We're already cyborgs—our phones are extensions of our minds. But 2.7 billion people remain in digital darkness. Dr Lollie Mancey, digital anthropologist and RTÉ Futureville co-presenter, challenges the notion that AI will free us to paint in meadows. Reality? Jobs will vanish, universal basic income may arrive, and we'll face a purpose crisis when work no longer defines us. She poses the era's defining question: when your AI assistant comes home, is it at your table or recharging in the garage? Your answer reveals how you see technology's role. She's betting 80% on AGI by 2030—not gradual progress, but desperate need for higher intelligence. From 1950s washing machines to ChatGPT, labour-saving tools never save time—they shift expectations. The future isn't written, and AI won't decide our fate—we will.Show NotesGuest: Dr Lollie ManceyTitle: Digital Anthropologist, Co-presenter of RTÉ's FuturevilleKey Topics:[01:46] Humans plus technology, not instead of - Why AI isn't pixie dust to sprinkle on everything. Anthropologists are finally having their moment as the human element becomes critical.[03:05] 2.7 billion in digital darkness - Who's missing from the AI conversation? Ireland's bubble makes us forget vast populations have no internet access.[04:23] Who benefits from time saved? - Will employers reward productivity over hours? The washing machine didn't free women from housework—it just changed expectations.[06:17] Universal basic income and purposelessness - When manual labour vanishes, what happens to identity? Two-generation unemployment creates malaise, addiction, depression. The pension (€270/week) is our only test case.[07:10] The 1970s leisure prediction - Someone walked into a Dublin classroom and wrote "leisure time vs work time" on the blackboard. Were they right? Will we choose to work, or will the choice be made for us?[26:31] AGI by 2030: 80% odds - Lollie's bold prediction: artificial general intelligence within six years, triggered by an unsolvable crisis requiring higher brain power.[27:21] The interruption problem - New voice AI that interrupts changes everything. If you're rude to ChatGPT, are you training yourself to be rude to humans?[28:08] At the table or in the garage? - The defining question: where does your AI companion belong? Younger generations already see them as household members, not machinery.[29:28] The invasive technology concern - Why Lollie and neuroscience professors agree: don't open the hard box protecting our soft brains unless absolutely necessary.Key Takeaways:We're passive cyborgs now; we need to become active by understanding algorithmsThe future of work isn't about productivity gains—it's about identity reconstructionAI adoption without considering the 2.7 billion offline is incomplete thinkingYour answer to "table or garage?" reveals your entire worldview on technologyLabour-saving tools historically shift work, they don't eliminate itResources:Futureville - RTÉ programme imagining Ireland in 2050Connect with Lollie: drlollie.ie | LinkedIn: Dr Lollie (L-O-L-L-I-E)Podcast: Available at drlollie.ieChatting GPT is produced by AI Institute. For AI adoption in built environment firms, visit https://weareaiinstitute.com/ Hosted on Acast. See acast.com/privacy for more information.

When Allan Russell joined Musgraves as Head of AI and Process Automation, he thought the 147-year-old Irish retailer wasn't using much AI. Within weeks of his "roadshow" around the business, he'd uncovered 17 unauthorised applications—before even speaking to IT. Then the real numbers emerged: 300,000 hits to Grammarly in 30 days across just 1,700 office users. Shadow AI wasn't theoretical—it was rampant and ungovernanced. Rather than panic, Allan built a strategy around what people were already doing, rolling out Microsoft Copilot and establishing four pillars: process automation, AI, culture, and governance. His advice? Stop hesitating—your competitors are already using these tools. And never deploy AI without first talking to the people whose problems you're trying to solve.Show NotesGuest: Allan Russell, Head of AI and Process Automation, MusgravesKey Topics:[04:16] The shadow AI discovery - How an informal "roadshow" uncovered 17 AI applications before consulting IT.[06:44] 300,000 Grammarly hits - Network data revealed 300,000 Grammarly hits in 30 days across 1,700 users, plus 1,200 ChatGPT hits.[08:02] Four foundational pillars - Process automation and AI supported by culture and governance. Why psychological safety matters as much as guardrails.[10:32] Rolling out Microsoft Copilot - Why starting with enterprise solutions beat playing whack-a-mole with shadow IT.[23:32] Two pieces of advice - (1) Don't hesitate—competitors are already moving. (2) Understand pain points first. Design thinking beats technology-first thinking.[24:57] Beyond internal consumers - What's next: recipe managers, price optimisation, identifying opportunities across Musgraves' franchisee ecosystem.Key Takeaways:Quantify your shadow AI problem through network traffic analysisRelationship capital precedes technical deploymentStart with pain points, not platformsGovernance enables speed by giving people permission to experiment safelyConnect: LinkedIn - Allan J RussellReplay of a 2024 episode. Allan's blueprint for managing shadow AI remains one of the clearest roadmaps we've covered. Hosted on Acast. See acast.com/privacy for more information.

When Munich's citizens were invited to reimagine one of their main streets using AI, the results didn't just sit in a presentation deck—they were actually built. Damiano Cerrone, founder of Urbanist AI and Co-Plan AI, has worked with over 80 organisations worldwide to transform how cities engage with their residents. Instead of traditional surveys asking "what do you want?", his platforms let people visualise, evaluate, and iterate on urban design ideas in real-time. The twist? Damiano argues AI models should be biased—tuned to each participant's perspective—because that's what makes collaboration meaningful. From training Eastern European mayors to test policies with generative AI, to working with the UN Development Programme and Dubai's Prime Minister's office, he's pioneering what he calls "collaborative public governance." The challenge ahead: making European city centres attractive to humans again after decades of infrastructure-led dehumanisation.Guest: Damiano CerroneCompany: Urbanist AI & Co-Plan AILocation: Helsinki, FinlandKey Topics Discussed[02:02] Munich's streets transformedHow a citizens' engagement process during a festival led to an AI-imagined green space and water feature being built as a prototype on one of Munich's main streets.[03:29] Training mayors in Eastern EuropeWorking with the UN Development Programme to help mayors use generative AI to evaluate and test their own policies—compressing months of work into rapid visualisation cycles.[05:33] The collaboration vs. replacement questionWhy cities are asking "why engage people when we can simulate engagement with AI agents?"—and Damiano's response: treating AI as a billion colleagues, not a billion replacements.[08:37] When what people want is wrongHow AI enables people to evaluate their own ideas by seeing them visualised, leading to minds being changed through the process itself rather than being told "no" by experts.[16:38] Embracing bias in AI modelsDamiano's contrarian stance: models should be tuned to individual participants' biases because that's what creates interesting, fruitful conversations—not sterile consensus.[18:13] The mega-trend for European citiesWhy attracting human beings back to dehumanised city centres is the biggest urban challenge ahead, following the post-COVID exodus.Resources MentionedUrbanist AI: site.urbanistai.comCityLab Berlin: Monthly neighbourhood engagement programme (Kids Lab)Futureville (RTÉ): Irish television series imagining Athlone in 2050Connect with Damiano: LinkedInFor Built Environment ProfessionalsDamiano's work demonstrates how participatory AI can close the gap between consultation and implementation—particularly relevant for architecture practices, engineering consultancies, and planning departments struggling with meaningful stakeholder engagement. His platforms have been adopted by governments worldwide, proving that citizen-generated spatial ideas can move from digital visualisation to physical intervention.Chatting GPT is produced by AI Institute. If you're interested in AI adoption for built environment firms, visit https://weareaiinstitute.com/ Hosted on Acast. See acast.com/privacy for more information.