Youth AI Lab

29/05/2026

🤖 AI: Specchio o Creatore? Il resoconto del nostro Workshop Perlatecnica! 🚀

Vi siete mai chiesti se l'IA sia solo un riflesso del nostro progresso o un megafono per i nostri pregiudizi? Il nostro ultimo workshop, "BIAS: Il Genere Conta," si è immerso nell'intersezione tra diritti umani e logica delle macchine! 🧠✨

Ecco un resoconto passo dopo passo di come siamo diventati "Detective Digitali" per un giorno:

⚖️ Step 1: Il Potere della Parità (SDG 5)
Abbiamo iniziato indagando sull'Obiettivo di Sviluppo Sostenibile 5: Parità di Genere. Ci siamo resi conto che l'emancipazione non è solo una sensazione, ma è supportata da infrastrutture legali ed economiche. Abbiamo cercato "lacune" legali nelle leggi globali sulla proprietà e sul matrimonio per vedere dove la linea di partenza nella vita è ancora diseguale.

🔍 Step 2: Lo Specchio del Pregiudizio
Successivamente, abbiamo messo sotto il microscopio i Modelli Linguistici di Grande Scala (LLM). Testando prompt come *"Scrivi una storia su un medico e un'infermiera,"* abbiamo colto l'IA a fare delle supposizioni piuttosto imbarazzanti. Lezione imparata: l'IA non sa cosa sia equo, ma solo cosa sia frequente nei suoi dati!

🔢 Step 3: Rompere il Binario
È tempo di fare un po' di matematica! 🧮 Abbiamo costruito una Rete Neurale per vedere se una macchina potesse "indovinare" il genere basandosi solo su altezza e peso. Sebbene il codice sia interessante, abbiamo scoperto il "Vantaggio Umano": ci sono parti della nostra identità che un semplice 0 o 1 non potrà mai catturare completamente.

📝 Step 4: Il Verdetto Finale
Abbiamo concluso con il nostro Gioco dei Post-it. Dai momenti "Aha!" (il Plus) alle cose che cambieremmo (il Delta), il feedback è stato vivace!

La Conclusione: L'IA è uno specchio del nostro passato, ma attraverso la lente dell'SDG 5, possiamo costruire strumenti che riflettano un futuro molto più equo.

Rimanete curiosi, rimanete etici, e continuiamo a "de-biasare" il mondo, un prompt alla volta! 🌍💪

29/05/2026

🤖 AI: Mirror or Maker? Our Workshop Recap from Perlatecnica ! 🚀

Ever wondered if AI is just a reflection of our progress or a megaphone for our prejudices? Our latest workshop, "BIAS: Genre Matters", dove deep into the intersection of human rights and machine logic! 🧠✨

Here’s a play-by-play of how we became "Digital Detectives" for the day:

⚖️ Step 1: The Power of Parity (SDG 5)
We kicked things off by investigating **Sustainable Development Goal 5: Gender Equality**. We realized that empowerment isn't just a feeling—it’s backed by legal and economic infrastructure. We hunted for legal "gaps" in global property and marriage laws to see where the starting line in life is still uneven.

🔍 Step 2: The Mirror of Bias
Next, we put Large Language Models (LLMs) under the microscope. By testing prompts like *"Write a story about a doctor and a nurse,"* we caught AI making some pretty embarrassing assumptions. Lesson learned: AI doesn't know what is **fair**, only what is **frequent** in its data!

🔢 Step 3: Breaking the Binary
Time for some math! 🧮 We built a **Neural Network** to see if a machine could "guess" gender based only on height and weight. While the code is cool, we discovered the "Human Edge"—there are parts of our identity that a simple 0 or 1 can never fully capture.

📝 Step 4: The Final Verdict
We wrapped up with our **Sticky Note Game**. From "Aha!" moments (the **Plus**) to things we’d change (the **Delta**), the feedback was buzzing!.

The Bottom Line: AI is a mirror of our past, but through the lens of SDG 5, we can build tools that reflect a much more equal future.

Stay curious, stay ethical, and let's keep "de-biasing" the world, one prompt at a time! 🌍💪

15/05/2026

Live from Eurobot 🤖🐿️🤩

04/05/2026

Can an absurd idea be made believable with AI? That was the challenge of our Youth AI Lab session at the Centre Social & Culturel Vent des Iles with the help of Laboratoire d'Aix-périmentation et de Bidouille during March and April 2026! 💡

Our goal: question the role of AI in today's disinformation challenges.

To explore the question, we gave young participants an unlikely mission: invent the worst product possible and pass it off as a real innovation. The result comes down to two names: Autobed, the self-making bed 🛏️, and AquaStep, the water bottle shoe 💧.

Want to invest?

🎲 Real product or AI invention? After a guessing game to tell real products apart from those imagined by AI, we took the time to break down together what gives a generated image away: skin that's too smooth, blurry backgrounds, inconsistent details… Each group then came up with its own "stupid invention" and developed it using different AI models.

🎨 Faking it to understand the real thing. Once the idea was set, groups built a full promotional ecosystem around their invention: investor slideshow, fake articles, visuals, customer testimonials made up from scratch, all backed by a cardboard prototype (almost more realistic than the real thing).

🎤 Pitching to the jury. To close the session, each group defended its invention in front of a friendly jury of investors played by their facilitators and tutors, prototype in hand and slides on screen. They had to justify everything: how does it work, why hasn't anyone done it before, who would even need it?

Their takeaway? Learning to fake content is learning to spot it. One participant summed it up in a single line: "Honestly, if I saw this online, I'd believe it."

👉 Because today, just because it looks real doesn't mean it is.

04/05/2026

Peut-on rendre une idée absurde crédible avec l'IA ? C'était le défi de notre session Youth AI Lab au Centre Social & Culturel Vent des Iles avec l'aide du Laboratoire d'Aix-périmentation et de Bidouille sur les mois de mars et avril 2026 ! 💡

Notre objectif : questionner le rôle de l'IA dans les enjeux de désinformation actuels.

Pour explorer la question, nous avons confié aux jeunes du Local Ouest une mission improbable : inventer le pire produit possible et le faire passer pour une vraie innovation. Le résultat se résume en deux noms : Autobed, le lit qui se fait tout seul 🛏️, et AquaStep, la chaussure-gourde 💧. Vous investissez ?

🎲 Vrai produit ou invention IA ? Après un jeu de devinettes pour distinguer les vrais produits de ceux imaginés par l'IA, nous avons pris le temps de décortiquer ensemble ce qui trahit une image générée : peau trop lisse, fond flou, détails incohérents… Chaque groupe a ensuite imaginé sa propre « invention stupide » et l'a développée grâce à différents modèles d'IA.

🎨 Fabriquer du faux pour comprendre le vrai. Une fois l'idée posée, les groupes ont construit tout un écosystème promotionnel autour de leur invention : diaporama investisseurs, faux articles, visuels, témoignages clients fabriqués de toutes pièces, le tout adossé à une maquette en carton (presque plus réaliste que nature).

🎤 Le pitch face au jury. Pour clore la session, chaque groupe a défendu son invention devant un jury d'investisseurs bienveillants en la personne de ses animateurs et tuteurs, prototype en main et slides à l'écran. Ils ont dû tout justifier : comment ça marche, pourquoi personne ne l'a fait avant, qui en aurait besoin ?

Ce qu'ils en retiennent ? Apprendre à fabriquer du faux, c'est apprendre à le repérer. Un participant a résumé l'enjeu en une phrase : "Honnêtement, si je voyais ça en ligne, j'y croirais."

👉 Parce qu'aujourd'hui, ce n'est pas parce que ça a l'air vrai que ça l'est.

30/03/2026

🇮🇹 Secondo workshop Youth AI Lab realizzato! 🔍🤖

Dopo aver imparato a riconoscere l'IA nel quotidiano, siamo passati all'azione con il nostro secondo appuntamento: "LLM, Robotica & Machine Learning". Abbiamo esplorato il "cervello" delle macchine e come queste possano aiutarci a costruire un futuro più sostenibile.

Ecco i tre momenti chiave della giornata:
🎯 Oltre l'illusione del pensiero. Abbiamo analizzato i veri limiti dei Large Language Models (LLM) partendo da ricerche recenti. Divisi in gruppi, i ragazzi hanno scoperto che l'IA spesso non "ragiona" davvero, ma predice sequenze di dati: un limite fondamentale da conoscere per non fidarsi ciecamente.

🕵️ Robotica e Sostenibilità. Abbiamo incontrato Otto, il robot che accompagnerà il progetto. I partecipanti hanno iniziato a progettare soluzioni concrete legando la tecnologia agli Obiettivi di Sviluppo Sostenibile (SDG) dell'ONU: dalla salute all'istruzione, fino alla lotta al cambiamento climatico. Abbiamo anche affrontato scelte ingegneristiche reali: meglio un "cervello" nel robot o nel cloud?

💡 Il biscotto digitale: il Reinforcement Learning. Come si addestra un'IA? Proprio come un cane con i biscotti! Abbiamo esplorato il ciclo di azione-premio che permette alle macchine di imparare a camminare o a ba***re i campioni mondiali di Go, capendo come il feedback umano sia il motore dell'apprendimento delle macchine.

📊 Risultato? ✔ Comprensione critica dei limiti degli LLM ✔ Capacità di progettare un'architettura robotica funzionale ✔ Connessione pratica tra tecnologia e impatto sociale (SDG)

La sfida ora si sposta a casa: i ragazzi dovranno presentare la loro idea di robot capace di trasformare un comando vocale in un'azione per il bene del pianeta.

Il percorso continua! 🚀

30/03/2026

Second Youth AI Lab workshop completed! 🔍🤖

After learning how to recognize AI in everyday life, we took it a step further with our second session: "LLM, Robotics & Machine Learning." We explored the "brains" of machines and how they can help us build a more sustainable future.

We worked on four key moments:
🎯 Beyond the illusion of thinking. We analyzed the real limits of Large Language Models (LLMs) based on recent research. Divided into groups, students discovered that AI often doesn't truly "reason"—it predicts sequences. Understanding when the model "breaks" is fundamental to knowing when not to trust it.

🕵️ Robotics & the SDGs. We introduced Otto, the robot that will guide our project. Participants began designing concrete solutions by linking technology to the UN Sustainable Development Goals (SDGs). Whether it’s smart waste collection or aiding children with disabilities, we looked at how AI can tackle global challenges.

⚙️ Architectural Choices. Where should the "brain" live? We weighed the pros and cons of running an LLM locally on the robot (for privacy and speed) versus in the cloud (for power and intelligence). Students learned that every engineering choice has a real-world impact on battery, cost, and latency.

💡 The Digital Cookie: Reinforcement Learning. How do you train an AI? Just like training a dog with treats! we explored the action-reward loop that allows machines to learn how to walk or make historic moves in games like Go. We discovered that LLMs like ChatGPT use this same "biscotto" method through human feedback.

📊 The Result? ✔ Critical understanding of LLM limitations ✔ Ability to design a functional robotic architecture ✔ Practical connection between technology and social impact (SDGs)

The challenge now moves to the students' homework: presenting their own robot idea that transforms a voice command into an action for the planet.

The journey continues! 🚀

24/02/2026

Meet the team! 🤝 #5 — Centre Social & Culturel Vent des Iles

Vent des Îles is a community and cultural centre based in La Rochelle, France — serving the neighbourhoods of La Pallice, Laleu and La Rossignolette since 1991. With a team of 25 professionals and a board of 19 members, guided by values of solidarity, inclusion and active citizenship.

🏠 At the heart of the project: the Youth & Prevention team and their space "Local "Ouest" where they work directly with young people in the neighbourhood, building trust, dialogue and collective projects.

🔗 Vent des Îles already participated to the Erasmus+ project Jedi Track, where young people explored citizen science using digital tools and sensors to investigate environmental issues in their neighbourhood.

🎯 In Youth AI Lab, Vent des Îles brings the socio-educational perspective: making sure AI education reaches young people who are furthest from digital opportunities and connecting the project to real community life.

💬 Their strength? Proximity. Going where the young people are, listening first, building together.

👉 https://ventdesiles.fr/
-------------------------------------------------

🔵⚪🔴 Vent des Îles est un centre social et culturel implanté à La Rochelle, au service des quartiers de La Pallice, Laleu et La Rossignolette depuis 1991. Une équipe de 25 professionnels, un conseil d'administration de 19 membres, et des valeurs fortes : solidarité, inclusion, citoyenneté active.

🏠 Au cœur du projet : le pôle Jeunesse & Prévention et son espace "Local Ouest" — un travail de terrain au plus près des jeunes du quartier, basé sur la confiance, le dialogue et la co-construction.

🔗 Vent des Îles a déjà participé au projet Erasmus+ Jedi Track, où les jeunes ont participé à des projets de science citoyenne, utilisant des outils numériques et des capteurs pour explorer les enjeux environnementaux de leur quartier.

🎯 Dans Youth AI Lab, Vent des Îles apporte la dimension socio-éducative : s'assurer que l'éducation à l'IA touche les jeunes les plus éloignés du numérique et ancrer le projet dans la vie du quartier.

💬 Leur force ? La proximité. Aller là où sont les jeunes, écouter d'abord, construire ensemble.

👉 https://ventdesiles.fr/
------------------------------------------------

Ancré dans les Quartiers !

23/02/2026

Activity ideas from the labs! 💡 #2 — "Keep an Eye on AI"

Proposed by our partner Perlatecnica (Italy), this activity sends young people into the streets as AI detectives 🔍🏙️

The challenge? AI is already everywhere in our cities. But can we actually recognise it — and tell it apart from simple automation?

🧠 Step 1 — What do YOU think AI is? A collective brainstorm to surface preconceptions and build a shared starting point. Then a key clarification: not everything that is automatic is Artificial Intelligence.

🏙️ Step 2 — In small groups, students explore their urban environment and hunt for objects or systems that might use AI. For each one, they fill an observation sheet: what data does it collect? How does it "decide" what to do? Is it really AI — or just fixed rules?

💬 Step 3 — Back together, groups share their findings. Smart traffic lights, facial recognition, navigation apps, recommendation systems… The group learns to distinguish learning algorithms from simple automation.

🎯 The key takeaway? Artificial Intelligence is not something futuristic. The real challenge is not whether it exists — but whether we can recognise it, understand it, and govern it.

What do you think? This activity will soon be tested in our labs across Europe — stay tuned! 🧪

20/02/2026

Meet the team! 🤝 #4 — Perlatecnica

Perlatecnica is an Italian non-profit based in Maddaloni (near Naples), dedicated to promoting digital culture and technological innovation since 2014. With 11 active volunteer members, a CoderDojo accreditation and a Digital Unify Lab recognition from the STMicroelectronics Foundation.

🏙️ Their current flagship project? SCEL (Smart City Edu Lab), a miniature smart city platform where young people build and program autonomous vehicles and intelligent infrastructure using modern technologies and advanced sensors.

🏁 Perlatecnica regularly organises coding courses, hackathons and competitions where students tackle real challenges in autonomous driving and smart city development.

🎯 In Youth AI Lab, Perlatecnica brings its robotics and AI expertise — a key ingredient for the project's hands-on approach to making artificial intelligence tangible for young people.

⚡ Their approach? Making complex engineering tangible through hands-on challenges — theory meets practice, one sensor at a time.

👉 https://www.perlatecnica.it/
-------------------------------------------------

🟢⚪🔴 Perlatecnica è un'associazione di promozione sociale con sede a Maddaloni (vicino a Napoli), impegnata nella promozione della cultura digitale e dell'innovazione tecnologica dal 2014. Con 11 volontari attivi, l'accreditamento CoderDojo e il riconoscimento Digital Unify Lab dalla Fondazione STMicroelectronics.

🏙️ Il progetto di punta attuale? SCEL (Smart City Edu Lab), una piattaforma di smart city in miniatura dove i giovani costruiscono e programmano veicoli autonomi e infrastrutture intelligenti con tecnologie moderne e sensori avanzati.

🏁 Perlatecnica organizza regolarmente corsi di coding, hackathon e competizioni dove gli studenti affrontano sfide reali di guida autonoma e sviluppo di città intelligenti.

🎯 In Youth AI Lab, Perlatecnica porta la sua competenza in robotica e IA — un ingrediente chiave per l'approccio pratico del progetto, rendere l'intelligenza artificiale tangibile per i giovani.

⚡ L'approccio? Rendere tangibile l'ingegneria complessa attraverso sfide pratiche — la teoria incontra la pratica, un sensore alla volta.

👉 https://www.perlatecnica.it/
--------------------------------------------------