The In-Between Tech and Trust Podcast

The Tech Naivety Undermining Democracy - Erdem Ovacik (EP 32)

🎙️ with Erdem Ovacik AI trust in our institutions is falling, and Erdem Ovacik argues it will keep falling. The technology is not what fails him here. The governance underneath it was never built to earn that trust, and pouring more AI into unfit democratic institutions is like adding fuel to a fire no one has learned to contain. This is for anyone watching their organisation or their government absorb technology faster than it can govern it. 🎧 Episode overview Ovacik has spent his career at the point where technology meets public life, and he starts from an uncomfortable claim: we are becoming more capable of trusting our institutions, and we won't. He walks through why representative democracy - one national representative for every fifty thousand people - concentrates power in a way that markets and modern tech companies solved for long ago, and asks what democracy would look like if it learned from them. Eva presses him on where technology solves the problem and where it amplifies it, on what he means by "tech naivety," and on the danger of handing decisions to an AI that becomes its own black box. The conversation sits with Europe's harder bind: build real technological competence, or live inside infrastructure and decisions shaped by those who did. It closes on what daily life might feel like if the next democracy actually arrived. 🔑 Key themes * Why adding more technology to a broken institution accelerates the loss of trust rather than repairing it * Whether a system built for fifty thousand people per representative can be governed at all in an age of AI * The line between AI as a consulted voice and AI as a new black box that decides for us * Europe's choice between building its own AI competence and depending on decisions made elsewhere * What it means to keep human agency when the machine could, in theory, write better policy than we can 👤 About the guest Erdem Ovacik is a social innovator, entrepreneur, and author of The Next Democracy. He built Donkey Republic, one of the first app-based bike-share systems, which is the detail that matters most here: he argues from inside the technology industry, not against it. His warning that technology "almost necessarily" harms society because addiction is good business carries weight precisely because he has spent years building the kind of product he now scrutinises. 🕒 Chapter markers * [01:22] Why we could trust our institutions more, and won't * [02:52] Fifty thousand to one: the power concentration inside representative democracy * [06:39] Tech naivety, and the fire you keep adding fuel to * [11:42] The handlebar problem: who sets the priorities, and where AI belongs * [13:35] Europe's bind, and what happens when the black box is built elsewhere * [17:48] What the next democracy would feel like to live in 🔗 Links * Erdem Ovacik on LinkedIn - https://www.linkedin.com/in/erdemovacik/ * Eva Simone Lihotzky on LinkedIn - https://www.linkedin.com/in/evalihotzky/ * The Next Democracy book - https://nextdemocracy.com * Related episode - Tech & Democracy and how both can be connected to create trust with Nexus Politics - https://open.spotify.com/episode/3WVnBsaz3LfFPwn6xsX1N6

Two Fears Behind Every AI Transformation - Dr. Charlotte Blum (EP 31)

🎙️ Dr. Charlotte Blum - Director Change & Organisational Design, edding Group Two fears run underneath every AI rollout, and most companies manage neither. Dr. Charlotte Blum works the human emotions of AI transformation - the fear, identity, and trust that decide whether adoption sticks - from a board-level role most organisations don't have. This is a conversation about why people resist AI, and why the technology itself may be the hardest thing in the room to trust. 📌 The episode Charlotte Blum was hired after an agility transformation failed, into a staff function reporting to the board, to do the thing change management usually skips: work with what people feel. In this conversation she separates two fears companies tend to blur - fear of the technology and fear of change itself - and argues that AI destabilises people at the level of identity, not workflow. She makes a case that AI is structurally untrustworthy for an unexpected reason: it has no self-interest. A system built only to please will give you the answer it thinks you want, not the true one. Asked whether she has seen an organisation genuinely succeed at AI transformation, her answer is short, and it isn't yes. For any leader who suspects the resistance in their team is emotional rather than logistical, this is a conversation about what the roadmap leaves out. 🧠 What you'll sit with * Why the fear of AI arrives before the technology does, and why it's measurable * Why an organisation that can't move someone's desk won't move them to AI * What happens to a person when their role, and their sense of self-worth, is suddenly in question * Why a tool built to please can't be relied on to tell you the truth * Why the thing to optimise for is trust, not results, not time 👤 About the guest Dr. Charlotte Blum is Director Change & Organisational Design at the edding Group, where she leads AI and organisational transformation from a rare board-facing position that treats the emotional side of change as core work rather than an afterthought. Her method is to enable leaders for the conversations no job description trained them for, and she has the internal data suggesting it moves the numbers. ⏱️ Chapters  [00:00] Trust too much, or not at all - the problem with both [02:16] Why AI is a human shift, not a technology one [07:12] The skill capitalism doesn't want you to build [13:48] Why leaders are handed a job they were never trained for [30:22] The trust equation, and where AI breaks it [38:11] Three things an organisation must get right 🔗 Links Dr. Charlotte Blum on LinkedIn - https://www.linkedin.com/in/dr-charlotte-blum/ [https://www.linkedin.com/in/dr-charlotte-blum/] edding Group - https://www.edding.com The trust equation (credibility + reliability + intimacy / self-orientation) - https://people-shift.com/articles/the-trust-equation/ AllBright Academy - https://www.allbright-stiftung.de/academy Related episode - Why Empathy Can't be Automated with Gifty Enright - https://open.spotify.com/episode/5NYYRgwROaiZLHur5gbksw

Tech and Trust at the Bundeswehr Cyber Innovation Hub: Innovating for and with the Troops with Sarah Marie Sandmann (EP 30)

🎙️ Sarah Marie Sandmann, Innovation & Intrapreneurship, Bundeswehr Cyber Innovation Hub AI trust in defence starts where slides end: with a soldier under pressure who needs to understand, rely on, and account for the technology in their hands. Sarah Marie Sandmann works at the Bundeswehr Cyber Innovation Hub in Innovation and Intrapreneurship, the official innovation unit of the German armed forces. Sandmann treats trust in defence technology as a capability criterion, something tested under pressure, not asserted in policy, and the organisations getting this right are rebuilding how innovation works within the institution itself. 📖 Episode overview The Bundeswehr is one of Europe's most structurally complex organisations, built for stability, accountability, and risk minimisation, not speed. Sandmann and her colleagues run innovation projects at 12-month cycles that would take years through standard procurement. This episode explores what that tension looks like in practice: how AI is deployed strictly as decision support rather than decision replacement, how soldiers co-develop the technologies they will eventually trust with their lives, and why a trustworthy defence innovation ecosystem would be measured by capabilities delivered rather than the quality of its presentations. Sandmann also reflects on the post-Ukraine shift she has observed from inside the institution — more civilians wanting to contribute, more startups engaging with defence, and what that change means for civil-military trust. 🔍 Key themes * Whether a soldier can understand, rely on, and explain an AI system, and why all three must be true before deployment * The structural case for why large institutions are slow to innovate, and why the people inside them usually aren't the problem * What "decision support, not decision replacement" means as a live design constraint for AI in high-stakes environments * How trust between military institutions and the startup ecosystem is actually built, and what breaks it * What a trustworthy defence innovation ecosystem would need to look like in two to three years 👤 About the guest Sarah Marie Sandmann works at the Bundeswehr Cyber Innovation Hub in Innovation and Intrapreneurship, the official innovation unit of the German armed forces. She works at the intersection of military capability development, startup collaboration, and responsible technology adoption, collaborating on projects that bring AI, autonomous systems, and emerging technologies into operational use through direct co-development with soldiers. She has been inside the institution through the post-Ukraine shift in civil-military engagement and speaks from that experience with unusual clarity. ⏱ Chapter markers * [00:00] What the Cyber Innovation Hub actually does — and why cockroaches are involved * [03:18] Trust as an operational requirement in defence technology * [08:00] Why innovation resistance is structural, not cultural * [13:09] AI as decision support — the bright line and how it holds * [21:22] The post-Ukraine shift and what a trustworthy ecosystem would look like 🔗 Links * Sarah Marie Sandmann on LinkedIn — https://www.linkedin.com/in/sarah-sandmann/ * Eva Simone Lihotzky on LinkedIn - https://www.linkedin.com/in/evalihotzky/ * Bundeswehr Cyber Innovation Hub — https://www.cyberinnovationhub.de/en/ * SwarmBioTactics and Autobugs project — https://www.youtube.com/watch?v=k4vu5AKTkJk [https://www.youtube.com/watch?v=k4vu5AKTkJk] * Komand.AI and Smart Lead project — https://www.youtube.com/watch?v=-r45um6txpQ [https://www.youtube.com/watch?v=-r45um6txpQ] * Sonic AI - https://www.youtube.com/watch?v=c9i98jrielw [https://www.youtube.com/watch?v=c9i98jrielw] * Related Episode: Why Security Intelligence Fails Before the Attack with Assaf Kipnies - https://open.spotify.com/episode/2D4ODAxGULFbqmXCmgwsfA

Anthropic's Model Suspension, Europe 2031, and G7 World Leaders Lunching with Frontier AI Lab & Tech CEOs: The Week in Tech & Trust with Yours Truly (EP 29)

🎙️ solo episode with host Eva Simone Lihotzky Anthropic's frontier AI model was pulled offline for every non-American in three days, and suddenly Europe's AI access looked less like something it owns and more like a permission. This is a week where digital trust stopped being abstract: one US export directive, one warning about Europe's compute future, and one lunch table where the people who build AI sat with the people who govern it. For any leader applying AI inside an organisation, it is a week worth understanding in practice, not as headlines. 🧭 In this episode In a single week of June 2026, three events landed that most coverage treated as separate. Eva reads them as one thread. The US Commerce Department forced Anthropic's Fable and Mythos models offline for any foreign national. The Europe 2031 agenda argued that Europe's window to matter in AI is closing faster than its own authors had predicted. And for the first time, Frontier AI lab CEOs sat at the G7 heads-of-state table. The question underneath all three: when access to the most strategic technology of the moment sits on someone else's permission, what does a European organisation actually own? Eva works through what this means for vendor dependency, infrastructure design, and the difference between treating AI sovereignty as a compute problem and treating it as a trust problem. 🔍 Key themes * Why "access" to a frontier AI model may be a permission that someone else can withdraw — and what that does to a strategy built on it * The gap between Europe's AI story as a capital problem and the trust assumption sitting underneath it * What changes for a leader when vendor lock-in stops being a risk slide and becomes a live event * Whether building infrastructure and orchestration across many models is now resilience rather than over-engineering * When the builders of AI also shape the rules that govern it, who represents the people using it 🎙️ About the host Eva Simone Lihotzky, AI adoption and ethics advisor, formerly MD in one of the largest independent agency groups in Europe and co-author of 10 Moral Questions: How to Design Tech & AI Responsibly. She has spent more than a decade leading AI implementation inside organisations, which is why this episode resists the easy reads — it stays with the gray zone between hypocrisy and conviction, between capital and trust, rather than resolving it. This is a solo reflection: Eva connecting three news events into one question she openly admits is hard to narrow down. ⏱️ Chapters * [00:00] Three news items, one thread * [02:22] A frontier model offline in three days * [08:00] Europe 2031: the window that closed early * [11:30] Mistral, and the scale of the gap * [18:45] The G7 table: builders meet the people who govern them * [25:10] Who represents the ones using the technology 🔗 Links * Eva Simone Lihotzky on LinkedIn — https://www.linkedin.com/in/evalihotzky/ * Europe 2031 agenda — https://europe2031.ai * 10 Moral Questions: How to Design Tech & AI Responsibly — https://www.10moralquestions.com/the-book * Eva's World Economic Forum reflection, January 2026 — The politics of tech on Spotify: https://open.spotify.com/episode/1RKtxdJWXcQH8vnpnDtgEP?si=wrln7peeSkKb-gotGHYMRg * Eva's World Economic Forum reflection, January 2026 — The politics of tech on Apple Podcasts: https://podcasts.apple.com/de/podcast/the-in-between-tech-and-trust-podcast/id1828521905?l=en-GB&i=1000747143762 * Anthropic statement on the Fable / Mythos suspension — https://www.anthropic.com/news/fable-mythos-access

AI and The Cognitive Atrophy Trap: What Happens When We Let Tech Shortcut the Hard Parts of Learning - Tobias Burkhardt (EP 28)

🎙️ Tobias Burkhardt, Founder of The Shift School AI, trust and learning are on a collision course, and the casualty is judgment. Tobias Burkhardt, founder of the Shiftschool, argues that the way individuals and organisations are adopting AI in learning is a cultural problem: the reflex to make learning faster and cheaper is precisely what makes AI dangerous to the people using it. This conversation is for anyone who suspects the upskilling programmes around them are solving for the wrong problem. 💡 Episode overview Tobias Burkhardt has spent years advising organisations on learning and organizational development, and his diagnosis is uncomfortable: cognitive atrophy is real, it is already happening, and it predates AI. The impulse to shortcut understanding — to reach for the tool before doing the thinking — is a cultural pattern that AI accelerates but did not create. In this conversation, he makes the case for treating AI as a relational technology rather than a productivity instrument, and for rebuilding learning around curation, community, and continuity rather than content delivery. He also names something most learning institutions will not: that the ultimate goal of good education is to make oneself obsolete. 🔑 Key themes * Why treating AI as a tool rather than a collaborator is ill-advised, and what the alternative requires * The faster-and-cheaper reflex in organisational learning, and why it compounds the problem it is meant to solve * What a school without content actually means, and what takes content's place * The bilateral responsibility in learning, and why self-discipline alone will never be sufficient * Trust as an investment: why waiting for certainty before engaging with AI is the wrong posture 🎤 About the guest Tobias Burkhardt is the founder of The Shiftschool, a learning institution he built because he loved learning and never liked schools. He advises organisations on learning strategies and has developed a philosophy of education built around what does not change — judgment, curation, social interaction, and continuity — rather than around the tools and content that do. His concept of a school without content is a practical response to the decreasing half-life of knowledge in an AI-native world. ⏱ Chapter markers * [00:00] Can we trust ourselves to use AI — not just trust AI itself * [04:00] Why the information abundance problem predates AI * [08:30] From tool to collaborator to environment — how the relationship with AI evolves * [11:00] Cognitive atrophy and the shortcutting reflex * [18:30] Lifelong learning as personal obligation — and why institutions cannot wait * [22:30] The school without content — what takes knowledge's place * [30:00] Redesigning Shift School for an AI-native world 🔗 Links * Tobias Burkhardt on LinkedIn: https://www.linkedin.com/in/meetropoly/ * Eva Lihotzky on LinkedIn: https://www.linkedin.com/in/evalihotzky/ * Visit the Shift School: https://shiftschool.de * Listen to the related episode with Simon Berkler on organisational AI adoption or trust in digital systems (EP 22): https://open.spotify.com/episode/6y8PMaVUnZVAR1hOAR15DN