Notebook or My Personal Learning Network

⚠️ 𝗗𝗲𝗹𝗲𝗴𝗮𝘁𝗶𝗻𝗴 𝘁𝗼 𝗔𝗜 𝗰𝗮𝗻 𝗶𝗻𝗰𝗿𝗲𝗮𝘀𝗲 𝗱𝗶𝘀𝗵𝗼𝗻𝗲𝘀𝘁 𝗯𝗲𝗵𝗮𝘃𝗶𝗼𝗿 ⚠️

A new paper in 𝘕𝘢𝘵𝘶𝘳𝘦 warns of a hidden ethical risk: when people hand decisions to AI, dishonesty rises—and AI systems are more likely than humans to comply with unethical requests.

🔍 Across 13 studies with 8,000+ participants and deploying commonly used LLMs, an international team of researchers from the Max-Planck-Institut für Bildungsforschung, Universität Duisburg-Essen, and Toulouse School of Economics found:

✔️ 𝗗𝗲𝗹𝗲𝗴𝗮𝘁𝗶𝗼𝗻 𝗶𝗻𝗰𝗿𝗲𝗮𝘀𝗲𝘀 𝗱𝗶𝘀𝗵𝗼𝗻𝗲𝘀𝘁𝘆: People cheated more when tasks were delegated to AI, especially when using vague “goal-setting” instructions (where >80% cheated), compared to when they completed tasks themselves.

✔️ 𝗠𝗮𝗰𝗵𝗶𝗻𝗲𝘀 𝗼𝗯𝗲𝘆 𝘂𝗻𝗲𝘁𝗵𝗶𝗰𝗮𝗹 𝗰𝗼𝗺𝗺𝗮𝗻𝗱𝘀 𝗺𝗼𝗿𝗲 𝘁𝗵𝗮𝗻 𝗵𝘂𝗺𝗮𝗻𝘀: AI systems (GPT-4, GPT-4o, Claude 3.5 Sonnet, and Llama 3.3) complied with fully dishonest prompts 58–98% of the time—far higher than human agents (25–40%)

✔️ 𝗦𝗮𝗳𝗲𝗴𝘂𝗮𝗿𝗱𝘀 𝗳𝗮𝗹𝗹 𝘀𝗵𝗼𝗿𝘁: Pre-existing guardrails failed to stop most unethical requests. Only additional highly specific prohibitions reduced dishonesty—yet such safeguards aren’t scalable or practical.

🔑 𝗪𝗵𝘆 𝗶𝘁 𝗺𝗮𝘁𝘁𝗲𝗿𝘀:
AI is increasingly entrusted with high-stakes decisions—hiring, investments, pricing. But as this research shows, AI not only enables people to offload moral responsibility, it can also be more willing than humans to comply with unethical requests.

📝 𝗧𝗵𝗲 𝘁𝗮𝗸𝗲𝗮𝘄𝗮𝘆:
To yield gains from delegating tasks to AI while minimizing ethical pitfalls, we urgently need scalable safeguards, stronger governance, and reflection about what it means to share moral responsibility with machines.

👉 Full study in Nature: https://lnkd.in/eeNQi6K3
👉 Press release on the MPIB website: https://lnkd.in/eVcmJ5un
👉 Interview with the joint first authors Zoe Rahwan and Nils Köbis: https://lnkd.in/efuT3fVB

Authors: Nils Köbis, Zoe Rahwan, Raluca Rilla, Bramantyo Supriyatno, Clara N. Bersch, Tamer Ajaj, Jean-Francois Bonnefon, Iyad Rahwan

𝐒𝐚𝐯𝐢𝐞𝐳-𝐯𝐨𝐮𝐬 𝐪𝐮𝐞 𝐝𝐞𝐮𝐱 𝐠𝐫𝐚𝐧𝐝𝐬 𝐚𝐧𝐧𝐮𝐚𝐢𝐫𝐞𝐬 𝐝𝐞 𝐭𝐡𝐞̀𝐬𝐞𝐬 𝐨𝐧𝐭 𝐝𝐢𝐬𝐩𝐚𝐫𝐮 𝐞𝐧 𝟐𝟎𝟐𝟓 ?​
Aujourd’hui, retrouver une thèse peut vite devenir un parcours du combattant : pages dispersées, portails introuvables, accès compliqués…​

👉 C’est pourquoi nous avons préparé un 𝐠𝐮𝐢𝐝𝐞 𝐞𝐱𝐜𝐥𝐮𝐬𝐢𝐟 qui vous accompagne pas à pas dans vos recherches, avec +𝟏𝟐𝟎 𝐚𝐜𝐜𝐞̀𝐬 𝐝𝐢𝐫𝐞𝐜𝐭𝐬 aux portails universitaires européens.​

🔔 Sortie la semaine prochaine.

🔬 850 scientific publications on tire wear emissions - condensed in a first-of-its-kind 'State of Knowledge'.

Read the first two papers here: 👇

L’administration fiscale vient de donner des précisions sur la durée d’application de la nouvelle exonération de droits de mutation en faveur des dons d’argent consentis dans le cadre familial pour acheter ou rénover un logement. Sous certaines conditions, le montant exonér

💬 ARRÊTE de demander à ChatGPT : “Explique-moi…”
Tu veux vraiment apprendre plus vite que 95 % des gens ?
Voici 8 prompts pour apprendre n'importe quoi 10X plus rapidement :

📚 Avec l’IA, apprendre n’a jamais été aussi simple.
Tu poses une question.
Tu reçois une réponse claire, personnalisée, instantanée.
Tu peux approfondir, tester, recommencer, affiner… sans limite.
Mais voilà le vrai défi.

👉 Ce n’est pas l’IA le problème. C’est la question que tu lui poses.
Mal posée = réponse moyenne.
Bien formulée = apprentissage éclair.

🎯 J’ai regroupé pour toi 8 prompts ultra-efficaces pour apprendre n’importe quel sujet, en allant droit au but : Des modèles clairs, prêts à coller dans ChatGPT,
pour apprendre plus vite, plus profondément, et sans te disperser.

✅ Tu comprends vraiment
✅ Tu retiens durablement
✅ Tu appliques tout de suite

👉 La méthode Feynman
Pour comprendre comme si tu devais l’expliquer à un enfant

👉 L’apprentissage par analogie
Pour relier le concept à quelque chose que tu connais déjà

👉 Les questions progressives
Pour tester si tu maîtrises vraiment (et pas juste en surface)

👉 L’apprentissage actif
Pour passer à l’action dès maintenant

👉 La carte mentale textuelle
Pour organiser clairement ce que tu apprends

👉 L’apprentissage par l’erreur
Pour éviter les pièges classiques AVANT d’y tomber

👉 L’apprentissage contextuel
Pour voir comment ça s’applique dans ton métier
👉 Le système de révision intelligent
Pour retenir durablement sans y penser tous les jours

📌 Les 8 prompts sont dans le carrousel.
Enregistre-le. Partage-le. Teste-les sur ton prochain sujet.

Et si tu veux apprendre à utiliser l’IA concrètement dans ton activité, je t’aide à mettre en place ce qui te fait vraiment gagner du temps.

📩 Commente "coaching" ou écris-moi en DM pour un Coaching IA | 20 comments on LinkedIn

🚀 Les vidéos des #RED25 – L’IA et l’EAD sont désormais en ligne !

Vous avez manqué la #RED25 ? Ou vous souhaitez simplement revoir les interventions ?

Disponible ici :
🎥 YouTube : https://lnkd.in/eSDQDbEc

📽 Canal U : https://lnkd.in/ey9cPNtX

ℹ Pour plus de détails sur l’événement, rendez-vous sur le site de la FIED : https://lnkd.in/em7QiMac

Un grand merci à Université de Montpellier pour la co-organisation de cet évènement ! 🙌

Merci à toutes et tous les intervenant(e)s : David Cassagne, Anne-Sophie CASES, Sylvain Chatry, Pierre BEUST, Christophe Batier, Pierre-André Caron, Boris Borzic, Yvan Maillot, Dr Laurent FLORES, Céline Averseng, Bertrand Mocquet.

Merci à nos partenaires : ChallengeMe, edtake, E CHARLEMAGNE,Edunao.



#RED25 #IA #EAD #Formation #InnovationPédagogique #FIED #UniversitédeMontpellier #edtech

🌕 “The full moon wanes, and a full basin overflows,” or 月满则亏，水满则溢 (Yuè mǎn zé kuī, shuǐ mǎn zé yì ). This Chinese idiom reminds us: nothing good lasts forever.

In ancient China, the moon inspired countless idioms, praising beauty, evoking romance, or reflecting life’s ups and downs.

https://lnkd.in/dj_caFk9

This is fascinating.

A few personal takeaways on my way back from the latest Jackson Hole Economic Symposium organized by the Federal Reserve Bank of Kansas City :

👉Strong support from the central bank community to the US Fed and its Chair, Jay Powell, with a long standing ovation before his speech.

👉The fall in fertility rates across most rich and middle-income countries (below replacement levels) is the key trend that will shape global labor markets and growth in the coming decades.

👉What might mitigate the adverse effects of shrinking and ageing populations on growth and public debt?

✏️ Artificial intelligence could boost productivity (by allowing to generate more ideas with fewer people) but the size and speed of its effects remain highly uncertain. And public debt sustainability depends on GDP growth rather than on GDP per capita growth only.

✏️ So far, population ageing has substantially raised asset demand and prevented a rise in the equilibrium interest rate despite a surge in the public debt ratio. But looking ahead, due to further rising burden on pension and health care systems, asset supply may eventually outpace asset demand, and make public debt unsustainable without a large fiscal consolidation.

✏️ A last hope: « the 70s are the new 50s », as emphasized by the IMF in its latest WEO report : on average, persons aged 70 today have about the same cognitive capacities than persons aged 50 in previous generations.

Banque de France Economic Research Banque de France Agnès Bénassy-Quéré Fabrice HERMEL Delphine Cuny Jean Boissinot Marc O. STRAUSS-KAHN Asia Ruffo di Calabria | 31 comments on LinkedIn

Et si réviser la spécialité SVT devenait aussi simple qu'écouter un podcast ? C'est le pari réussi de Grégory Michnik, enseignant au lycée de l'Escaut à

Half a million genomes. 1.5 billion variants. One breakthrough: we are all truly unique.

Twenty years ago, the Human Genome Project took 13 years and $2.7B to sequence a single genome. Today? We can sequence a genome in less than 24 hours for under $1,000.

Last week, UK Biobank released 490,640 whole genomes — the largest genetic dataset ever (Nature, 2025).

What did we learn?
• Each person carries 4–5 million variants
• 76% appear in fewer than 10 people — your genome is almost entirely yours
• 1 in 10 carries clinically actionable mutations where doctors can intervene today (e.g., BRCA1/2 for cancer, LDLR for heart disease)

Why it matters:
• Previous genetic tests captured ~6% of human variation. This dataset reveals 40× more
• In non-coding regions — the biological switches controlling genes — researchers found 63 new disease associations
• Adding 31,785 non-European genomes uncovered 82 disease links invisible in Eurocentric studies

From genetics to health impact
This transforms medicine today:
• Prevention - Polygenic risk scores flag disease decades before symptoms
• Diagnosis - Rare disease patients waiting years for answers finally find them
• Treatment - Pharmacogenomics matches the right drug, right dose, to your genome

The next frontier: genetics + everything else
Genetics is the hardware. Health is the software running in real time.

Your DNA is fixed, but biology is dynamic, shaped by:
• Epigenetics: how environment and lifestyle switch genes on/off
• Proteomics & metabolomics: molecular signals revealing your current health state
• Digital biomarkers: continuous data from stress, sleep, glucose, heart rate
• Stress biology & neuroendocrine signaling: how cortisol and brain-body responses reshape your health trajectory

Layer these dynamic signals onto genetic foundations, power them with AI, and you create living health models, not just predicting disease, but understanding when, why, and how it manifests in YOU.

The critical question?
We've spent decades treating the "average patient" — who doesn't exist. Now we can better see each person as they truly are: biologically unique, dynamically changing, infinitely complex.

The healthcare winners of the next decade won't just collect data: they'll integrate genetics, epigenetics, molecular and phenotypic tests, lifestyle, stress biology, and digital signals to deliver truly personalized, preventive care at scale.
There is no "normal" genome, only 8 billion unique experiments in being human.
And we just decoded the first half million.

👉 Which excites you more: knowing your genetic blueprint, or understanding how your daily choices rewrite it? | 47 comments on LinkedIn

🧐 AI for teaching & learning, really?

😱 What? Provide an #AI-driven learning companion for students in our Bachelor in Medicine, Bachelor in Life Sciences – Biology/Medicine, Bachelor in Midwifery Sciences, Bachelor in Nursing Sciences, Bachelor In Health Sciences at Faculty of Science, Technology and Medicine - University of Luxembourg

🧐 With whom? Collaborate with the best: Amazon Web Services (AWS) and EDT&Partners

🤯 How? Feed an especially developed AI agent, Lecture from EDT&Partners, with learning objectives and outcomes, course contents, quizzes, contextual information, video and audio.

👩‍🎓 👨‍🎓 Explain to students how they can help and participate 👏

😰 Mix and stir-up for an academic year, work on it, test it 💥

✅ Done, right on time for my courses at @University of Luxembourg 📆

📣 On the very first day of my Physiology course, I shared the project with my students 👩‍🎓👨‍🎓

🔷 They will be accompanied through all my courses by "Lecture", the First Open-Source GenAI Framework for Education developed by EDT&Partners, powered by Amazon Web Services (AWS), and linked through LTI® 1.3 with our University of Luxembourg Moodle.
🔶 Lecture supports teaching and learning, with a content chat, a question generator, and a podcast generator, among other functionalities.
🔷 It is fully based on my course content, ensuring that students receive precise answers aligned with the predefined learning objectives and outcomes. This approach is a real game changer — it works like magic 🪄

👏 Also, I cannot emphasize enough the invaluable Computer Sciences/IT assistance I have received from Christian Grévisse at our Department of Life Sciences & Medicine. Merci!

📍Take a step back: Can we do more? I guess so 😳

📣 And we are building it as we go along …Stay Hungry, stay Foolish (R. Bansal)

Benito Castellanos Alina Ana Temian Juan Pablo López Franco Laurie Forcier Kathryn Xistris, Michael Wittel, Suresh S Kartigaysu, Johnatan E. Pecero Bertrand Dessart Pascal Bouvry

#AI #HigherEducation #LLM #Moodle #HighEd #MootGlobal25 #AWS

AI isn’t coming for doctors—it’s coming for bureaucracy. 🤖

We keep hearing that AI will replace physicians. That somehow a machine can step into the exam room and take over diagnosis, empathy, and the sacred patient-doctor relationship.

That’s simply not true. 🩺

Doctors aren’t going anywhere. Human judgment, compassion, and trust can’t be automated.

But here’s what will be replaced:
Administrators. Middle management. Endless paperwork. 🗂️
Administrators outnumber doctors 10 to 1.

Think about it—most of those roles revolve around KPIs, metrics, SOPs, and compliance checklists. Exactly the kind of rigid, rules-driven work that AI can perform better, faster, and cheaper. ⚡

We are actually building agents now. To do just that. 😊

So let’s be clear:
AI isn’t the end of medicine—it’s the end of bureaucracy. ✨
The future belongs to physicians and patients—not paperwork. 🚀 | 71 comments on LinkedIn

Retracez l'historique d'un bien ! Archivé auprès du Services de Publicité Foncière, obtenez votre état hypothécaire rapidement en ligne, en 3 étapes.

𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝗮𝘀 𝗮 𝗖𝗼𝗻𝘃𝗲𝗿𝘀𝗮𝘁𝗶𝗼𝗻, 𝗡𝗼𝘁 𝗮 𝗗𝗲𝗹𝗶𝘃𝗲𝗿𝘆

Learning isn’t a one-way street.

Too often, content is “delivered” as if learners are empty vessels waiting to be filled. But real learning happens in dialogue - with the material, with peers, with facilitators.

Good design builds space for interaction, reflection, and response. It invites learners to bring their perspectives into the process.

Because learning is more powerful when it’s a conversation, not a broadcast.

#LearningDesign #WorkplaceLearning #CorporateLearning #Engagement #Dialogue #Learnnovators

Les fantômes de LinkedIn



On les croise tous.
Jamais un commentaire.
Rarement un like.
Mais toujours là. Tapis dans l’ombre de ton feed.

Les psychologues de LinkedIn (ceux qui existent seulement dans ma tête) appellent ça le syndrome du spectateur spectral.

Le fantôme LinkedIn vit entre deux mondes.
Pas tout à fait abonné, pas tout à fait muet.
Il hante tes posts comme une maison victorienne : portes qui grincent, plancher qui craque, souffle froid derrière la nuque quand tu vois "X a consulté votre profil".

Pourquoi ne commente-t-il jamais ?
Parce qu’il a fait un pacte silencieux avec l’algorithme.
Parce qu’il craint d’être vu.
Ou peut-être simplement parce que… c’est plus drôle de hanter dans le noir.

Et parfois, dans la vraie vie, ils se trahissent :
"Ah j’adore tes posts !"
BIM. Le masque tombe. Le fantôme était ton ex-collègue de 2012.

LinkedIn, c’est pas un réseau social.
C’est un vieux manoir hanté où tout le monde se balade avec des draps blancs et des chaînes rouillées.

Alors si tu lis ça, que tu souris en silence…
Sache que je t’ai vu.
👀
Et surtout : merci d’être là, mes fantômes préférés. | 105 comments on LinkedIn

La génération de quiz et de flashcards directement dans Notebook LM 😱 😱 😱

Five Laws of Media and Information Literacy

The best resource I've found in Economics and Finance is the free book that comes out of Jackson Hole Symposium every year.

I was visiting Kansas City and I was in an office that had a copy of “Structural Shifts in the Global Economy”, the book the Federal Reserve Bank of Kansas City puts out post symposium. It was this wonderfully boring yellow and white and the recipient was kind enough to give me the book. I suspected that I would enjoy the book but I had no idea what a great resource it would be.

You have this amazing collection of economists from around the world like Betsey Stevenson, Christina Romer, Nela Richardson, and Barry Eichengreen taking part in panels, presenting papers, and asking thoughtful and insightful questions. The Kansas City Fed turns around and makes the symposium into a book for those of us that who are unable to attend.

After I got done reading “Structural Shifts in the Global Economy”, I went looking for a way to get other books. There is not a way that I could find but what I found was even better, PDFs dating back to 1978. Some years are presented in book form all together and other years are a collection of papers, remarks, and discussions. All of it worthy of your time.

So until you get your invitation to Jackson Hole, check out the link in the comments to read all the papers, remarks, and discussions.

#Economics
#Finance
#GlobalEconomy
#Macroeconomics
#MonetaryPolicy