Sept. 17, 2026

The Algorithm Doesn't Speak Italian: AI in Italy: What We're Using, and What Can't Be Automated

The Algorithm Doesn't Speak Italian: AI in Italy: What We're Using, and What Can't Be Automated
Spotify podcast player icon Spotify Spotify
Spotify podcast player icon Spotify Spotify
Season 2

We knew that by the time we recorded, edited and released this episode, things would have advanced even further and sure enough they have. While we are currently seeing tech bros lament the too-fast progress of the machines they created, here us regular humans are trying to wade through what to think about it all...

but what about Italy?

Italy was the first country in the EU to pass a comprehensive AI law. It ranks seventh in the world for AI research, and yet, fewer than one in ten of its small businesses has started a single AI project.

Into that contradiction, Valentina brings something more personal: a newsletter she wrote about FOBO — the fear of becoming obsolete — after the same question landed twice in one week, once from a table of fellow freelancers all twenty years into their careers, and once from her own AI students, decades younger and already asking if they'll have jobs at all.

Joining Georgette and Valentina is the formidable Nadia Koski, a global marketing and communications leader specializing in AI transformation and host of the new podcast Still Human: Real Talk in the Age of AI. The conversation moves fast: why entry- and mid-level jobs are the ones actually at risk right now, the two-to-three-year lag before global AI trends land in Italy, real (and not-so-reassuring) numbers on how Italians are using these tools without understanding them, Italy's new AI law and a deepfake incident that hit disturbingly close to home for Valentina's own family, a classroom study with a genuinely counterintuitive result, and the income and gender gaps already baked into who benefits from AI. It closes with Cose a Caso — including the AI use nobody wants to admit to, and the one Italian job that may be truly automation-proof.

Episode highlights

  • Where Valentina's FOBO newsletter actually came from: a table of fellow 40-something freelancers (including Giulia Scarpaleggia of Giulia's Kitchen) all asking the same question her own AI students ask her
  • Nadia on why entry-level and mid-level jobs face the first wave of disruption — with Meta's 10% layoffs and Klarna's 30%-cut-then-rehire as live examples
  • Valentina's "line": delegate everything to AI and you fall under it; use it to amplify your own judgment and you stay above it
  • "AI is always on LSD" — Valentina's classroom trick for teaching students to fact-check everything a chatbot gives them
  • Why Valentina tells students not to trust AI as a research tool, and the shift away from a straightforward Google search bar
  • The "washing machine effect" — running the same prompt through ChatGPT, Gemini, and Claude to compare results
  • Real Italian adoption numbers: 77% of Italians say they use AI tools (up from 69% in 2024), but only 7% claim in-depth understanding
  • Italy's AI law in practice — prison terms for harmful AI-generated content, and the deepfake incident involving Valentina's son's schoolmate that nearly went to the police
  • A study comparing three groups of high schoolers' AI use: the unrestricted-access group scored worst on tests; the group with guided, integrated use scored best
  • The US comparison: the Take It Down Act, deepfake victims, and who's already using the law to their own advantage
  • The World Bank stat Nadia cites: large language models are trained on roughly half the world's data — meaning half the world is missing from the information shaping all of this
  • SME AI adoption in Italy has doubled to 16% (from just over 8%) but still sits at half the EU average, with only 29% of small businesses applying AI to their core business
  • How AI is reshaping content and photography budgets for small businesses — and the much bigger disruption already hitting fashion e-commerce photo teams (see verification note)
  • Nadia on the one thing she still won't hand to AI: her own voice, her own writing, and who she trusts to help her with the rest
  • Cose a Caso: an AI palm reading, a chatbot relationship-advice confession, and the one Italian job *hint: it involves judgmental older men at construction sites* that nobody thinks AI can take

Guest

Nadia Koski is a global marketing and communications leader specializing in AI transformation, and the host of Still Human: Real Talk in the Age of AI — a new podcast examining technology's intersection with human experience, work, and identity. Based in Italy, she works globally, bringing a perspective that's genuinely international and unwilling to just cheerlead for the technology she works in. She previously created the Women Lead podcast and co-built The Digital Distillery podcast network, and now guest-lectures at Syracuse University, Florida State University, and the Florence School of the Arts. Based in Florence for the past 20 years, she also serves as a United Nations Women's delegate. [🔗 Still Human on Apple Podcasts · LinkedIn

If this episode resonated with you, leave us a review, drop a comment, and share it with someone who needs to hear it.

Where to find us

YouTube · Spotify · Apple Podcasts · Instagram

Special thanks to Vivace Media for producing us and Zoworking for letting us record our magic at your studio!