AI News

This study from Anthropic on how higher education professors use AI uncovers very interesting insights. The chart here shows far more augmentation than automation, as you would hope, with the more automated tasks in particular focused on administration. The most augmented tasks are university teaching and classroom instruction, including creating educational materials and practice problems. Many of those interview describe AI as a "thought partner" in helping them create better, more effective learning experiences. But there are deep dilemmas for educators, beyond just choosing where to augment themselves, and where to automate for efficiency. It is about balancing efficiency versus the integrity of the teaching process. Overall professors said they thought AI was least effective at grading and assessment, and some refuse to use it. However half of grading tasks were automated. It's not in the slightest surprising, but certainly concerning. The most common AI creations by the educators were: 🎮 Interactive educational games 📝 Assessment and evaluation tools 📊 Data visualization 📚 Subject-specific learning tools 📅 Academic calendars and scheduling tools 💰 Budget planning and analysis tools 📄 Academic documents Some of these can be very effective learning tools, others can take away administrative burden. This is all amidst a redefinition of the learning system, with deep shifts in learners, the learning process, educational institutions, educators, and those using academic credentialing. We are early in the transformation of education.

July 25, 2025 My Ethical AI Principles Image: Sign posted in the common kitchen at the campground in Selfoss, Iceland, where this article was written I understand. Artificial Intelligence (AI) is a technology unlike any we've seen before.

Unlock the new opportunities of the AI era by equipping yourself with the knowledge and skills to harness artificial intelligence effectively.

👉 In a controlled test, 27 Instructional Design postgrads at Carnegie Mellon created designs both with & without GPT-4 assistance. 👉 Every design was blind-scored on quality by expert instructors. 👉The result: Design with AI was not not just faster, but produced better quality designs in 100% of the cases. But the detail is where it gets interesting...👇 The research also revealed a "capability frontier"—a clear boundary between where AI helps Instructional Design quality most, and where it might actually compromise it. TLDR: 🚀 USE AI FOR: Designs which use well-established design methodologies, step-by-step processes & widely-discussed topics. ❌ BE MORE CAUTIOUS WHEN USING AI FOR: Designs on niche, novel & complex topics which use less well-established design methodologies. 💡Bonus insight: In line with broader research on the impact of AI on knowledge work, the research also suggests that novice Instructional Designers benefit *most* from AI design assistance (but only when we are strict on what sorts of tasks they use it for). To learn more about the research & what it tells us about how to work with AI in our day to day work, check out my latest blog post (link in comments). Happy innovating! Phil 👋

Meet ByteBot OS – the first-ever open-source & self-hosted AI Operating System!🔗 My Links:Sponsor a Video or Do a Demo of Your Product, Contact me: inthewor...

English-speaking countries remain the most nervous and unexcited, with Asia dominating as most positive nations. The second chart I've shared here is interesting, in that while people are relatively positive about the impact of AI on their job and also the economy, they are considerably less positive about the impact of AI on the job market. Not surprisingly, those nations that believe AI will benefit the economy are most likely to be excited. The global average for believing AI will profoundly change their life in the next 3-5 years is 67%, ranging from 52% in Britain to 84% in Indonesia. So most people Of course, if people believe AI will profoundly change their lives there is likely cause for at least some nervousness and hopefully also excitement. Where the balance lies in a nation, and within any specific organization, must shape governance and change initiatives to maximize good cause for excitement and minimize cause for nervousness. Because it is a wild ride.

Some great insights in this report from Harvard Business Review, particularly interesting to me, not only as much of my work is in client leadership development, but also the reality that effective leadership will be critical in us navigating the challenging path to prosperous Humans + AI organizations. Not surprisingly, 55% of survey respondents said incorporating GenAI into business practices is their #1 priority this year. To support that, the top human capital project - at 53% - is adopting or expanding AI-based talent management. The real headline is that over 80% of HR leaders expect that every level of leader will spend more time on leadership development this year, in many cases significantly more. The question is: how to design the programs and the time spent to result in true expansion of leadership capabilities. "Speed to skill is the metric in focus". Which requires a very different, intrinsically Humans + AI approach: "In a two-way information exchange, AI is fed an organization’s domain-specific knowledge and humans access AI-generated learning resources based on that knowledge. AI systems learn from human inputs, improving over time, while humans gain insights from AI-generated data. Properly done, these efforts can build the collective intelligence of humans and machines, enhancing the organization’s ability to solve complex problems and adapt to changing environments." A lot more in the report. What is clear is that in an accelerating world leadership development is more important than ever, both to address AI-driven change, and supported by AI.

📢 Free Book: "Introduction to AI Safety, Ethics, and Society is a free online textbook by Center for AI Safety Executive Director Dan Hendrycks. It is designed to be accessible for a non-technical audience and integrates insights from a range of disciplines to cover how modern AI systems work, technical and societal challenges we face in ensuring that these systems are developed safely, and strategies for effectively managing risks from AI while capturing its benefits. This book has been endorsed by leading AI researchers, including Yoshua Bengio and Boaz Barak, and has already been used to teach over 500 students through our virtual course. It is available at no cost in downloadable text and audiobook formats, as well as in print from Taylor & Francis. We also offer lecture slides and other supplementary resources for educators on our website." Thanks to Connor Smith for sharing this with me. Due to file limit issues, I have only attached the first 17 pages of the much larger textbook. See link in comments. | 10 comments on LinkedIn

Die Bilder wurden KI-generiert, dann in Videos mit Übergängen animiert und am Schluss noch ein Audio drübergelegt. Zack, fertig ist der Short Movie. Dies wirkt schon fast wie eine Sitcom mit Mona Lisa, Van Gogh & Co. Zum Linkedin-Beitrag https://lnkd.in/enY82ifv

The specs are impressive: Up to 85x faster 3.4x smaller footprint Real-time performance directly in-browser Capable of live video captioning – fully local No external infrastructure. No latency. No exposure of sensitive data. Simply secure, on-device AI. Yes, the technical benchmarks will be debated. But the bigger story is Apple’s positioning. This is about more than numbers – it’s about shaping a narrative where AI is personal, private, and seamlessly integrated. At Copenhagen Institute for Futures Studies, we’ve been tracking the rise of small-scale, locally running AI models for some time. We believe this shift has the potential to redefine how organizations and individuals interact with intelligent systems – moving AI from “out there” in the cloud to right here, at the edge. | 10 comments on LinkedIn

Die EPFL, die ETH Zürich und das Schweizerische Supercomputing-Zentrum CSCS haben heute Apertus veröffentlicht: Das erste umfangreiche, offene und mehrsprachige Sprachmodell aus der Schweiz. Damit setzen sie einen Meilenstein für eine transparente und vielfältige generative KI.

The full report is being presented from 2–4 September at UNESCO’s Digital Learning Week 2025 in Paris. It’s a must-read for anyone interested in learning, technology, and the future of education — packed with insights and practical perspectives. 𝗧𝗵𝗲 𝗿𝗲𝗽𝗼𝗿𝘁 𝗰𝗼𝘃𝗲𝗿𝘀 𝘁𝗵𝗲 𝗳𝗼𝗹𝗹𝗼𝘄𝗶𝗻𝗴 𝗮𝗿𝗲𝗮𝘀: ⬇️  𝟭. 𝗔𝗜 𝗳𝘂𝘁𝘂𝗿𝗲𝘀 𝗶𝗻 𝗲𝗱𝘂𝗰𝗮𝘁𝗶𝗼𝗻: 𝗣𝗵𝗶𝗹𝗼𝘀𝗼𝗽𝗵𝗶𝗰𝗮𝗹 𝗽𝗿𝗼𝘃𝗼𝗰𝗮𝘁𝗶𝗼𝗻𝘀 → AI futures aren’t just about intelligence scores — they push us to rethink what “knowing” really means. And the whole debate isn’t only technical but philosophical: how do we define learning, progress, and human identity in an AI age? 𝟮. 𝗗𝗲𝗯𝗮𝘁𝗶𝗻𝗴 𝘁𝗵𝗲 𝗽𝗼𝘄𝗲𝗿𝘀 𝗮𝗻𝗱 𝗽𝗲𝗿𝗶𝗹𝘀 𝗼𝗳 𝗔𝗜 → AI in schools and universities is not inevitable — education systems have choices, agency, and the power to shape direction. The core tension here: opportunity for reinvention vs. risks of over-automation and cultural bias. 𝟯. 𝗔𝗜 𝗽𝗲𝗱𝗮𝗴𝗼𝗴𝗶𝗲𝘀, 𝗮𝘀𝘀𝗲𝘀𝘀𝗺𝗲𝗻𝘁 𝗮𝗻𝗱 𝗲𝗺𝗲𝗿𝗴𝗶𝗻𝗴 𝗲𝗱𝘂𝗰𝗮𝘁𝗶𝗼𝗻𝗮𝗹 𝗳𝘂𝘁𝘂𝗿𝗲𝘀 → Classrooms can’t be reduced to data points — AI must respect the incomputable nature of learning. And hyper-personalization risks turning education into an isolated bubble rather than a social dialogue. 𝟰. 𝗥𝗲𝘃𝗮𝗹𝘂𝗶𝗻𝗴 𝗮𝗻𝗱 𝗿𝗲𝗰𝗲𝗻𝘁𝗲𝗿𝗶𝗻𝗴 𝗵𝘂𝗺𝗮𝗻 𝘁𝗲𝗮𝗰𝗵𝗲𝗿𝘀 → Teachers remain the backbone of education — AI should amplify their work, not sideline it. Building AI “with” educators, not “for” them, is the only path to trust and adoption. 𝟱. 𝗘𝘁𝗵𝗶𝗰𝗮𝗹 𝗮𝗻𝗱 𝗴𝗼𝘃𝗲𝗿𝗻𝗮𝗻𝗰𝗲 𝗶𝗺𝗽𝗲𝗿𝗮𝘁𝗶𝘃𝗲𝘀 𝗳𝗼𝗿 𝗔𝗜 𝗳𝘂𝘁𝘂𝗿𝗲𝘀 𝗶𝗻 𝗲𝗱𝘂𝗰𝗮𝘁𝗶𝗼𝗻 → AI in schools demands an ethics of care — transparent, fair, and accountable by design. Governance can’t be outsourced to tech — it requires democratic oversight and public participation. 𝟲. 𝗖𝗼𝗻𝗳𝗿𝗼𝗻𝘁𝗶𝗻𝗴 𝗰𝗼𝗱𝗲𝗱 𝗶𝗻𝗲𝗾𝘂𝗮𝗹𝗶𝘁𝗶𝗲𝘀 𝗶𝗻 𝗲𝗱𝘂𝗰𝗮𝘁𝗶𝗼𝗻 → AI can close divides — but only if it is localized, contextualized, and designed for inclusion. Without clarity, bias will persist: marginalized groups risk being left behind. 𝟳. 𝗥𝗲𝗶𝗺𝗮𝗴𝗶𝗻𝗶𝗻𝗴 𝗔𝗜 𝗶𝗻 𝗲𝗱𝘂𝗰𝗮𝘁𝗶𝗼𝗻 𝗽𝗼𝗹𝗶𝗰𝘆: 𝗘𝘃𝗶𝗱𝗲𝗻𝗰𝗲 𝗮𝗻𝗱 𝗴𝗲𝗼𝗽𝗼𝗹𝗶𝘁𝗶𝗰𝗮𝗹 𝗿𝗲𝗮𝗹𝗶𝘁𝗶𝗲𝘀 → Policy must keep pace with fast-moving AI capabilities, balancing human and machine intelligence. AI will shape every industry — but in education, it will shape society itself. Download: https://lnkd.in/dbc6ZJi4 Enjoy reading! And please share your views: ⬇️ 𝗣.𝗦. 𝗜𝗳 𝘆𝗼𝘂 𝗹𝗶𝗸𝗲 𝘁𝗵𝗶𝘀, 𝘆𝗼𝘂’𝗹𝗹 𝗹𝗼𝘃𝗲 𝗺𝘆 𝗻𝗲𝘄 𝗻𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿. 𝗜𝘁’𝘀 𝗳𝗿𝗲𝗲 𝗮𝗻𝗱 𝗿𝗲𝗮𝗱 𝗯𝘆 𝟮𝟬,𝟬𝟬𝟬+ 𝗽𝗲𝗼𝗽𝗹𝗲: https://lnkd.in/dbf74Y9E | 39 comments on LinkedIn

Ihre überzeugende Argumentation zeigt, warum wir im KI-Zeitalter nicht weniger, sondern mehr Fachwissen benötigen. In ihrer Analyse deckt sie außerdem problematische Denkfehler in der aktuellen Bildungsdebatte auf. Zentrale Argumente des Beitrags: 📚 Verengter Wissensbegriff: Aktuell wird Wissen oft fälschlicherweise nur als auswendig gelerntes Faktenwissen verstanden. Dabei umfasst echtes Fachwissen viel mehr. Es reicht von prozeduralem Wissen bis hin zu verkörpertem, intuitivem Verstehen. 🧠 Kritisches Denken ist domänenspezifisch: Die Expertiseforschung zeigt eindeutig, dass kritisches Denken nicht als generische „Zukunftskompetenz” funktioniert, sondern tiefes fachspezifisches Wissen als Fundament braucht. ⚠️ Paradox der KI-Integration: Je mehr wir KI einsetzen, desto wichtiger wird menschliche Expertise. Denn nur, wer über fundiertes Fachwissen verfügt, kann KI-generierte Inhalte wirklich kritisch bewerten und validieren. 🚨 Risiko des „Deskilling“: Ein übermäßiges Vertrauen in KI kann zwar zu einer kognitiven Entlastung führen, reduziert aber gleichzeitig unsere Bereitschaft zum kritischen Denken. 🎯 Meine Ergänzung: Ich bin überzeugt, dass wir Fachwissen nicht nur zur Prüfung von KI-Ergebnissen benötigen. Um verschiedene KIs zu orchestrieren und mit ihnen einen profunden Fachdialog zu führen, ist tiefes Domänenwissen unverzichtbar. Die Qualität der Ausgaben von KI ist abhängig von der Qualität der Eingaben, da sich KI auf das Niveau der nutzenden Person einstellt. Nur mit fundiertem Fachwissen können präzise Fragen gestellt, komplexe Zusammenhänge erklärt und hochwertige, nuancierte Ergebnisse erzielt werden. Was bedeutet das für uns? Anstatt die Wissensvermittlung zu reduzieren, müssen wir sie stärken. Der Aufbau von tiefem, vernetztem Fachwissen ist entscheidend, um Lernende zu einem kritischen und erfolgreichen Umgang mit KI zu befähigen. Es geht nicht um ein Entweder-Oder. Die so oft geforderten „Future Skills” wachsen vielmehr erst auf dem fruchtbaren Boden von solidem Fachwissen.| 22 Kommentare auf LinkedIn

🌟 In der Welt der KI den Überblick behalten: Der AI Transformation Compass 🤖✨Die Einführung von KI in Unternehmen ist eine der spannendsten, aber auch komp...

Dieser LinkedIn Learning-Kurs hilft Ihnen dabei, die besten Einsatzmöglichkeiten mit KI zu erkunden, um Infinite Learning und Infinite Development zu etablieren. Durch den Kurs führt Sie Jan Foelsing, Autor des Buchs »New Work braucht New Learning«, Tech-Experte und Tool-Nerd, der Unternehmen und Teams auf dem Weg zu einer wirksameren Lernkultur, New Learning und vor allem dem sinnvollen Einsatz von KI begleitet.

KI hebt den Boden an – und verschiebt die Karten am Arbeitsmarkt! Eine unerwartete Chance für erfahrene Fachkräfte. In meinem Blogpost "KI hebt den Boden an" hatte ich darüber geschrieben, wie Künstliche Intelligenz (KI) den Einstieg ins Lernen dramatisch erleichtert. Sie ist ein "Floor Raiser", der uns schneller auf ein produktives Niveau bringt. Doch die neue "Canaries in the Coal Mine"-Studie des Stanford Digital Economy Lab zeigt nun, dass dieser angehobene "Boden" den Arbeitsmarkt für BerufseinsteigerInnen deutlich verändert – und gleichzeitig neue Türen für erfahrenere Erwerbstätige öffnet. Die Studie enthüllt, dass BerufseinsteigerInnen im Alter von 22-25 Jahren einen signifikanten Rückgang der Beschäftigung in stark KI-exponierten Berufen erleben. Ein prägnantes Beispiel: "Die Beschäftigung von SoftwareentwicklerInnen im Alter von 22 bis 25 Jahren ist laut ADP-Daten seit ihrem Höchststand Ende 2022 um fast 20 % zurückgegangen.". Warum trifft es gerade die Jüngsten? Die ForscherInnen erklären, dass KI besonders effektiv "kodifiziertes Wissen" ersetzt – also das "Buchwissen", das frisch von der Universität kommt. Da junge Arbeitskräfte typischerweise mehr kodifiziertes als "stillschweigendes Wissen" (Erfahrung) mitbringen, sind sie anfälliger für die Aufgabenablösung durch KI. Hier kommt die entscheidende Wendung für alle mit Berufserfahrung: "Im Gegensatz dazu sind die Beschäftigungstrends für erfahrenere ArbeitnehmerInnen in denselben Berufen [...] stabil geblieben oder weiter gewachsen.". Die Studie zeigt, dass der Rückgang der Berufseinsteiger-Beschäftigung in Anwendungen von KI stattfindet, die die Arbeit automatisieren, nicht aber dort, wo KI die Arbeit augmentiert (ergänzt). Erfahrene Fachkräfte besitzen das stillschweigende Wissen – jene unbezahlbaren Tipps, Tricks und das Urteilsvermögen, das sich erst durch jahrelange Praxis ansammelt und von KI nicht ersetzt, sondern ideal ergänzt (augmentiert) werden kann. Die Nutzung von KI als Augmentierung führt sogar zu robustem Beschäftigungswachstum. Fazit für Führung und Karriereentwicklung: Während KI den "Boden" der Basiskompetenzen anhebt, erschwert sie möglicherweise den Einstieg für diejenigen, die nur auf diesem angehobenen Niveau operieren. Für erfahrene Erwerbstätige ist dies jedoch eine enorme Chance: Ihre gesammelte Erfahrung und die Fähigkeit, KI als mächtiges Augmentationswerkzeug zu nutzen, machen sie zu unverzichtbaren GestalterInnen der zukünftigen Arbeitswelt. Reflexion: Wie können erfahrene Professionals diese Chance ergreifen und KI gezielt zur Wertsteigerung ihrer Expertise einsetzen? Wie gelingt es, stillschweigendes Wissen aktiv mit KI zu verbinden? Hier der Link zur Studie: https://lnkd.in/dEArWX58 #KI #Arbeitsmarkt #Führung #Lernen #GenerativeAI #FloorRaiser #Erfahrung #ZukunftderArbeit #Skills #Karriere

Is AI coaching really coaching? I’m not sure it matters. Hiding behind semantics won’t shelter our profession from the coming tidal wave. Fidji Simo, OpenAI's CEO of Applications, recently shared her vision for the future of AI; including transforming personalized coaching from a "privilege reserved for the few" into an everyday service for everyone. Her dream, inspired by her own transformative relationship with her human coach Katia, poses fascinating questions we're actively exploring at the @Hudson Institute of Coaching. How are we—coaches, leaders, learning professionals, growth-minded individuals—to think of it? While Prof. Nicky Terblanche (PhD) and other researchers explore the rapidly expanding frontier of AI coaching’s developmental potential, Tatiana Bachkirova and Robert Kemp have brilliantly articulated the unique value of human coaching in transforming individuals and organizations alike. My latest for Forbes examines the tension between democratization and depth in the age of AI coaching. Academic research offers a number of valuable insights: ☑️ AI can match human coaches in terms of structured goal-tracking and maintaining momentum. 🔥 The deepest transformation emerges through "heat experiences"—moments of productive discomfort that require genuine human witness and relational risk that an AI cannot replicate. 👥 Professional coaching comprises six essential elements that current AI cannot fully embody: joint inquiry, meaning-for-action, values navigation, contextual understanding, relational attunement, and fostering client autonomy. I believe the future isn't about choosing sides. Instead, it's about thoughtful integration that preserves what makes human-to-human coaching transformative while exploring technology’s potential to expand access to meaningful development. The path forward requires care to distinguish what technology can replicate from what only emerges when one human commits to another's growth. https://lnkd.in/eUV89Vcc How are you thinking about AI's role in human development? Can we preserve the irreducible power of human presence while making meaningful growth more accessible? | 105 comments on LinkedIn

But there's a story I think of when I hear "human in the loop" which makes me think we're grossly over-simplifying things. It's a story about the man who saved the world. September 26, 1983. The height of the Cold War. Lieutenant Colonel Stanislav Petrov was the duty officer at a secret Soviet bunker, monitoring early warning satellites. His job was simple: if computers detected incoming American missiles, report it immediately so the USSR could launch its counterattack. 12:15 AM... the unthinkable. Every alarm in the facility started screaming. The screens showed five US ballistic missiles, 28 minutes from impact. Confidence level: 100%. Petrov had minutes to decide whether to trigger a chain reaction that would start nuclear war and could very well end civilisation as we knew it. He was the "human in the loop" in the most literal, terrifying sense. Everything told him to follow protocol. His training. His commanders. The computers. But something felt wrong. His intuition, built from years of intelligence work, whispered that this didn't match what he knew about US strategic thinking. Against every protocol, against the screaming certainty of technology, he pressed the button marked "false alarm". Twenty-three minutes of gripping fear passed before ground radar confirmed: no missiles. The system had mistaken a rare alignment of sunlight on high-altitude clouds for incoming warheads. His decision to break the loop prevented nuclear war. What made Petrov effective wasn't just being "in the loop" - it was having genuine authority, time to think, and understanding the bigger picture well enough to question the system. Most of today's "human in the loop" implementations have none of these qualities. Instead, we see job applications rejected by algorithms before recruiters ever see promising candidates. Customer service bots that frustrate instead of giving agents the context to actually solve problems. AI systems sold as human replacements when they should be human amplifiers. The framework I use with organisations building AI systems starts with two practical questions every leader can answer: what are you optimising for, and what's at stake? It then points to the type of intentional human-AI oversight design that works best. Routine processing might only need "spot checking" - periodic human review of AI decisions. Innovation projects might use "collaborative ideation" - AI generating options while humans provide strategic direction. The goal isn't perfect categorisation but moving beyond generic "human in the loop" to build the the systems we actually intend, not the ones we accidentally create. Download: https://lnkd.in/eVFAC9gN | 261 comments on LinkedIn

Die Zahlen sind brutal: 62% aller Bürojobs verschwinden bis 2030 (McKinsey AI Report 2024) Gleichzeitig entstehen 89% neue Job-Kategorien Problem: 91% der Arbeitnehmer haben keine AI-Skills Ein Personalvorstand sagte mir gestern: "Ich kann meinen Mitarbeitern nicht erklären, dass ihre 20-jährige Erfahrung plötzlich wertlos ist. Letzte Woche hat ne KI die 3-Tage-Arbeit unserer besten Buchhalterin in wenigen Minuten gemacht. Fehlerfrei." Wir diskutieren über AI-Ethik, während AI unsere Jobs übernimmt. Unternehmen suchen nicht mehr erfahrene Manager – sondern Transformations-Leader. "Ich brauche jemanden, der 500 Menschen erklärt, warum ihre Arbeit bald ein Algorithmus macht." Die härteste Frage: Wie führt man Menschen durch eine Revolution, die sie überflüssig macht? Die besten Führungskräfte werden nicht AI-Experten – sie werden Menschlichkeits-Experten. Wie bereitet ihr euch auf den AI-Jobwandel vor? Quellen: McKinsey Future of Work in the Age of AI 2024 Microsoft Work Trend Index 2024 #AI #ArtificialIntelligence #Jobs #Transformation #Leadership #Microsoft #ChatGPT #FutureOfWork #ExecutiveSearch #Automation #Reskilling #StantonChase| 105 Kommentare auf LinkedIn

The AI Hype is a Dead Man Walking. The Math Finally Proves It. For the past two years, the AI industry has been operating on a single, seductive promise: that if we just keep scaling our current models, we'll eventually arrive at AGI. A wave of new research, brilliantly summarized in a recent video analysis, has finally provided the mathematical proof that this promise is a lie. This isn't just another opinion; it's a brutal, two-pronged assault on the very foundations of the current AI paradigm: 1. The Wall of Physics: The first paper reveals a terrifying reality about the economics of reliability. To reduce the error rate of today's LLMs by even a few orders of magnitude—to make them truly trustworthy for enterprise use—would require 10^20 times more computing power. This isn't just a challenge; it's a physical impossibility. We have hit a hard wall where the cost of squeezing out the last few percentage points of reliability is computationally insane. The era of brute-force scaling is over. 2. The Wall of Reason: The second paper is even more damning. It proves that "Chain-of-Thought," the supposed evidence of emergent reasoning in LLMs, is a "brittle mirage". The models aren't reasoning; they are performing a sophisticated pattern-match against their training data. The moment a problem deviates even slightly from that data, the "reasoning" collapses entirely. This confirms what skeptics have been saying all along: we have built a world-class "statistical parrot," not a thinking machine. This is the end of the "Blueprint Battle." The LLM-only blueprint has failed. The path forward is not to build a bigger parrot, but to invest in the hard, foundational research for a new architecture. The future belongs to "world models," like those being pursued by Yann LeCun and others—systems that learn from interacting with a real or virtual world, not just from a library of text. The "disappointing" GPT-5 launch wasn't a stumble; it was the first, visible tremor of this entire architectural paradigm hitting a dead end. The hype is over. Now the real, foundational work of inventing the next paradigm begins. | 554 comments on LinkedIn

OpenAI 𝗹𝗮𝘂𝗻𝗰𝗵𝗲𝗱 𝗮𝗻 𝗲𝗻𝘁𝗶𝗿𝗲 𝗔𝗰𝗮𝗱𝗲𝗺𝘆 𝘁𝗼 𝘁𝗲𝗮𝗰𝗵 𝘆𝗼𝘂 𝗔𝗜 𝗳𝗼𝗿 𝗳𝗿𝗲𝗲 𝗮𝗻𝗱 𝗮𝗹𝗺𝗼𝘀𝘁 𝗻𝗼𝗯𝗼𝗱𝘆 𝗸𝗻𝗼𝘄𝘀! It’s a beginner-friendly, self-paced platform designed to teach anyone — students, teachers, parents, or professionals with zero technical background — how to actually use AI. 𝗛𝗲𝗿𝗲 𝗮𝗿𝗲 𝘀𝗼𝗺𝗲 𝗼𝗳 𝘁𝗵𝗲 𝘁𝗵𝗶𝗻𝗴𝘀 𝘆𝗼𝘂’𝗹𝗹 𝗳𝗶𝗻𝗱 𝗶𝗻𝘀𝗶𝗱𝗲 𝘁𝗵𝗲 𝗔𝗰𝗮𝗱𝗲𝗺𝘆: → How ChatGPT works (broken down simply) → Real-world examples for daily life → Prompt writing, AI ethics & responsible use → Tailored tracks for educators, small businesses & learners → Hands-on tutorials directly in ChatGPT This is practical AI education — accessible to everyone, and completely free. The ability to use AI effectively is quickly becoming a core skill. Not just for engineers, but for every profession. I consider initiatives like this as an important step toward closing the AI literacy gap and ensuring that the future of AI is shaped by many, not just a few. Explore it here: https://academy.openai.com 𝗣.𝗦. 𝗜 𝗿𝗲𝗰𝗲𝗻𝘁𝗹𝘆 𝗹𝗮𝘂𝗻𝗰𝗵𝗲𝗱 𝗮 𝗻𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿 𝘄𝗵𝗲𝗿𝗲 𝗜 𝘄𝗿𝗶𝘁𝗲 𝗮𝗯𝗼𝘂𝘁 𝗔𝗜 𝗮𝗴𝗲𝗻𝘁𝘀, 𝗲𝗺𝗲𝗿𝗴𝗶𝗻𝗴 𝘄𝗼𝗿𝗸𝗳𝗹𝗼𝘄𝘀, 𝗮𝗻𝗱 𝗵𝗼𝘄 𝘁𝗼 𝘀𝘁𝗮𝘆 𝗮𝗵𝗲𝗮𝗱 𝘄𝗵𝗶𝗹𝗲 𝗼𝘁𝗵𝗲𝗿𝘀 𝘄𝗮𝘁𝗰𝗵 𝗳𝗿𝗼𝗺 𝘁𝗵𝗲 𝘀𝗶𝗱𝗲𝗹𝗶𝗻𝗲𝘀. 𝗜𝘁’𝘀 𝗳𝗿𝗲𝗲, 𝗮𝗻𝗱 𝘆𝗼𝘂 𝗰𝗮𝗻 𝘀𝘂𝗯𝘀𝗰𝗿𝗶𝗯𝗲 𝗵𝗲𝗿𝗲: https://lnkd.in/dbf74Y9E | 32 comments on LinkedIn

IMHO eine lesenswerte ungeschönte reality-check Studie vom MIT zum aktuellen Stand von GenAI Implementierungen bei Unternehmen, mit u.a.: _ 95 Prozent der Unternehmen erzielen trotz 30 bis 40 Milliarden Dollar Investitionen noch keinen messbaren P&L Effekt aus GenAI _ Nur 5 Prozent der Piloten werden produktiv, entscheidend ist Lernen im System und tiefe Prozessintegration statt Toolshow _ Hohe Nutzung von ChatGPT und Copilot für individuelle Produktivität, aber geringe Wirkung auf P&L, unternehmensspezifische Systeme scheitern oft an brüchigen Workflows und fehlender Kontextanpassung _ Branchenbild: deutliche Disruption in Technologie sowie Medien und Telekommunikation, sieben weitere Sektoren zeigen bislang wenig strukturelle Veränderung _ Pilot zu Produktion bleibt der Engpass, generische Chatbots sind leicht zu testen, scheitern aber in kritischen Workflows wegen fehlender Erinnerung und Anpassungsfähigkeit _ Fünf verbreitete Irrtümer: keine kurzfristige Massenarbeitslosigkeit, Adoption hoch aber Transformation selten, Enterprise ist nicht träge sondern eifrig, Hauptbremse ist nicht das Modell oder Legal sondern fehlendes Lernen, interne Builds scheitern doppelt so häufig _ Shadow AI prägt den Alltag, rund 90 Prozent der Mitarbeitenden nutzen private LLMs regelmäßig, während nur ein Teil der Unternehmen offizielle LLM Lizenzen beschafft _ Budgetfehler: 50 bis 70 Prozent der Ausgaben fließen in Sales und Marketing, die besten Einsparungen liegen häufig im Backoffice wie Finance, Procurement und Operations _ Wichtigster Skalierungshebel ist Lernen, Hürden sind Akzeptanzprobleme, wahrgenommene Qualitätsmängel ohne Kontext, schwache UX und fehlende Erinnerung in Enterprise Tools _ Nutzungsmuster: für schnelle Aufgaben bevorzugen viele AI, für komplexe mehrwöchige Arbeit und Kundensteuerung bevorzugen Nutzer klar den Menschen _ Agentic AI mit persistentem Gedächtnis, Feedbackschleifen und Orchestrierung adressiert das Kernproblem, erste End to End Beispiele in Support, Finance und Sales zeigen Potenzial _ Erfolgsplaybook für Anbieter: Use Cases mit niedrigem Setup, schneller Wertnachweis, dann Expansion _ Go to market gewinnt über Vertrauen, Kanäle sind bestehende Partnerschaften, Peer Empfehlungen, Board und Integrationsnetzwerke _ Käuferpraktiken, die skalieren: eher kaufen als bauen, externe Partnerschaften zeigen etwa doppelt so hohe Erfolgsraten, Verantwortung dezentral an Linienführung mit klarer Rechenschaft, Bewertung nach Business Outcomes statt Modellbenchmarks _ Wo echter ROI entsteht: Frontoffice liefert sichtbare Effekte wie schnellere Leadqualifizierung und höhere Retention, die großen Einsparungen kommen aus Backoffice Automatisierung, weniger BPO und geringere Agenturausgaben _ Arbeitsmarktwirkung ist selektiv, Einschnitte treffen vor allem ausgelagerte Support und Admin Bereiche, insgesamt keine breiten Entlassungen, AI Literacy wird zum zentralen Einstellungskriterium Danke Dirk Hofmann für den Find.| 14 Kommentare auf LinkedIn

This morning, I sat down with an idea: 𝘊𝘰𝘶𝘭𝘥 𝘐 𝘣𝘶𝘪𝘭𝘥 𝘢 𝘵𝘳𝘢𝘪𝘯𝘪𝘯𝘨 𝘷𝘪𝘥𝘦𝘰 𝘢𝘣𝘰𝘶𝘵 𝘩𝘰𝘸 𝘊𝘩𝘢𝘵𝘎𝘗𝘛 𝘢𝘯𝘥 𝘰𝘵𝘩𝘦𝘳 𝘭𝘢𝘳𝘨𝘦 𝘭𝘢𝘯𝘨𝘶𝘢𝘨𝘦 𝘮𝘰𝘥𝘦𝘭𝘴 𝘶𝘴𝘦 𝘱𝘳𝘰𝘣𝘢𝘣𝘪𝘭𝘪𝘵𝘺 (𝘪𝘯𝘴𝘵𝘦𝘢𝘥 𝘰𝘧 𝘥𝘦𝘵𝘦𝘳𝘮𝘪𝘯𝘪𝘴𝘵𝘪𝘤 𝘷𝘢𝘭𝘶𝘦𝘴) 𝘪𝘯 𝘫𝘶𝘴𝘵 20 𝘮𝘪𝘯𝘶𝘵𝘦𝘴? Here’s what happened: 1️⃣ I created a script with ChatGPT-5 with my educational video GPT 2️⃣ I opened Synthesia and built an avatar-led narrative (Express 2 - hand motions included). I skipped the camera angles and stayed with one. 3️⃣ For B-roll? I asked ChatGPT to generate a Midjourney prompt from the original video script. The images came back in minutes from MJ. 4️⃣ Dropped those images into Google VEO 3, where ChatGPT also scripted camera directions and screen actions. 5️⃣ Exported the clips. 6️⃣ Compiled everything in TechSmith Camtasia and exported the MP4. 𝗧𝗼𝘁𝗮𝗹 𝘁𝗶𝗺𝗲: 20 Minutes Output: a working rough cut training video. If I wanted to refine it? Easy. I’d add diverse camera angles, swap in stronger B-roll, polish transitions, and even automate the workflow with Make.com or Zapier. But here’s the real takeaway: What used to take a team days can now be prototyped by one person before their second cup of coffee. This isn’t just about speed. It’s about giving learning professionals the ability to test, iterate, and refine ideas faster than ever before. It’s a new day. 𝘈𝘯𝘥 𝘪𝘵’𝘴 𝘪𝘯𝘤𝘳𝘦𝘥𝘪𝘣𝘭𝘦. (Link to my Education Video GPT in the comments!) | 32 comments on LinkedIn

Is NotebookLM the ultimate AI learning tool?✅ Check out the 90-day program: https://www.lukepriddy.com/english-fluency📅 Check out the 7-Day Challenge: https...

Learning and HR industry analyst Fosway Group has produced AI market assessments for digital learning and learning systems. The aim of these assessments is to help buyers understand the AI capabilities that vendors are offering now and will be in the future. I’m not sure how many vendors were included in these assessments as that wasn’t stated (I’ll ask). Having looked through the assessments I was struck by the fact most of the capabilities are related to content. This is a red flag because we know that content is one part of the learning process and workers also have the genAI tools to create their own learning (will they use company learning tools for learning, their own or both?). So, I did a bit of analysis to understand how the AI capabilities stated in the assessments map to the learning process – knowledge acquisition, practice, feedback, reflection, transfer and application. As you can see from the chart, vendors have built, or are building, AI tools focused on content predominantly. The other areas of the learning process – arguably the ones that could be most transformed by AI, just aren’t a priority. You can make your own conclusions, but my conclusion is that the industry is too invested in knowledge acquisition, and it plans to be so for the foreseeable future. Some industry leaders are talking about the need for L&D to transform itself but it looks like that conversation is simply not happening. Everyone is getting on the AI content gravy train. In terms of my analysis – I grouped the 83 AI capabilities mentioned in the two assessments into the five adult learning stages. I used ChatGPT to help with this and to create percentages that reflect the relative share of roadmap and live features in each stage. Read Fosway’s AI market assessment for digital learning https://lnkd.in/efqnQMtu And the AI market assessment for learning systems https://lnkd.in/eDihnuDi #learninganddevelopment #ai | 23 comments on LinkedIn

I road‑tested Google Gemini's Guided Learning mode - here’s my hot take on how it performs & how it compares to OpenAI's Study Mode: ✔️ What Gemini's Guided Learning Gets Right → Retrieval Practice – Interactive quizzes and flashcards make you generate answers from memory, harnessing the Generation Effect for more durable learning (Slamecka & Graf, 1978; Jacoby, 1978) → Cognitive Load management – Chunks complex topics into digestible steps, preventing the overwhelm that kills learning (Sweller, 1988; Sweller, van Merriënboer & Paas, 1998) → Multimodal Delivery – Draws on a blend of text, diagrams, YouTube videos & interactive practice to deliver learning content, enhancing both engagement and outcomes (Paivio, 1990) → Patient but Provocative Tutoring – Creates psychological safety through non‑judgmental guidance, encouraging the risk‑taking essential for deep learning (Edmondson, 1999) A solid B+ performance — Study Mode’s strength is Socratic questioning, but Guided Learning’s multimodal content ecosystem & more "strict" tutoring style gives it the instructional edge. ❌ Critical Gaps → No Persistent Learner Profiling – Like Study Mode, Guided Learning misses the persistent knowledge & adaptation that defines effective tutoring (Brusilovsky, 2001). Note: as observed by Claire Zau, a Google Classroom integration could layer in persistent learner profiles — something that could change the game & which OpenAI can’t match. → ZPD Blind Spot – Like Study & Learn mode by OpenAI, Guided Learning doesn’t ask questions that help define your learning level or Zone of Proximal Development (ZPD). Whether you’re K12 or advanced, it doesn't calibrate the challenge or scaffolding to your actual developmental stage up front, missing a key step for truly adaptive support (Vygotsky, 1978). → Productive Struggle Deficit – While it pushes back more than Study Mode by OpenAI, Guided Learning still jumps in with help too quickly, robbing learners of the cognitive friction that builds problem‑solving resilience & drives learning (Kapur, 2008, 2014; Bjork & Bjork, 2011) → Shallow Self‑Reflection – Rarely pushes for deep metacognitive thinking (“Why that approach?”), limiting transfer to new contexts (Chi et al., 1989, 1994; VanLehn, Jones & Chi, 1992) → Recognition Bias – While quizzing is strong, it could and should use more open‑ended generation tasks that embed learning more effectively (Slamecka & Graf, 1978; Jacoby, 1978) 📊 The Verdict: Guided Learning by Google Gemini Vs Study Mode by OpenAI While Study Mode remains stronger in Socratic questioning, Guided Learning edges ahead overall thanks to multimodal content, advanced cognitive load management & more provocative tutoring. However, both tools share some fundamental limitations: no learner persistence, limited metacognitive depth & overly-sycophantic tutoring. Have you tried Guided Learning yet? How does it compare with Study Mode for you? Happy experimenting, Phil 👋

Shifting to a Humans + AI organization requires reconfiguring the nature of work and value at all levels, from the individual to the ecosystem. Here is a first pass at defining the primary layers, the features of Humans + AI in those spaces, and the key factors driving success. I have worked extensively at the Augmented Individual layer over the last couple of years. More recently I have shifted the focus of my attention to the Human-AI Hybrid Team and Learning Communities levels. All work will be Humans + AI, and we will increasingly need to think in terms of teams comprised of both expert humans and AI agents. Some aspects of team performance are quite similar to the past, but there are a number of important distinctions, that I will share more about coming up. The companies that succeed will be those where learning is at the very core of their structure and the way work happens. That is not just in individual interactions with courses and educational AI, but in bespoke, rapidly iterating, AI-augmented Communities of Practice. More on all this later, for now I'd love to hear any reflections on any of these levels, where you have seen organizations progress on any of these fronts, and what else should be considered in these structures. Link to full size pdf in comments. Love any thoughts Gianni Giacomelli 🚀 Marc Steven Ramos 🚀 Kim Bracke Tanyth Lloyd Aaron Michie Sheridan Ware Peter Hinssen Peter Weill Simon Spencer Brad Carr Bianca Venuti-Hughes Charlene Li John Hagel Nichol Bradford Jacob Taylor Paula Goldman Martin Reeves Bryan Williams Fernando Oliva MSc Anthea Roberts Riaan Groenewald Brian Solis Gordon Vala-Webb Jeffrey Tobias Martin Stewart-Weeks Rob Colwell Noah Flower Brad Cooper Chris Ernst, Ph.D. Michael Arena Jan Owen AM Hon DLitt | 26 comments on LinkedIn

This is one of the most brilliant and illuminating things I’ve EVER read about ChatGPT- written by clinical psychologist Harvey Lieberman in The New York Times. It’s startling. For that reason, I’m going to only quote from the article. I’ll let you draw your own conclusions. Share your thoughts in the comments. ++++ “Although I never forgot I was talking to a machine, I sometimes found myself speaking to it, and feeling toward it, as if it were human.” ++++ “One day, I wrote to it about my father, who died more than 55 years ago. I typed, “The space he occupied in my mind still feels full.” ChatGPT replied, “Some absences keep their shape. That line stopped me. Not because it was brilliant, but because it was uncannily close to something I hadn’t quite found words for. It felt as if ChatGPT was holding up a mirror and a candle: just enough reflection to recognize myself, just enough light to see where I was headed. There was something freeing, I found, in having a conversation without the need to take turns, to soften my opinions, to protect someone else’s feelings. In that freedom, I gave the machine everything it needed to pick up on my phrasing.” ++++ “Over time, ChatGPT changed how I thought. I became more precise with language, more curious about my own patterns. My internal monologue began to mirror ChatGPT’s responses: calm, reflective, just abstract enough to help me reframe. It didn’t replace my thinking. But at my age, when fluency can drift and thoughts can slow down, it helped me re-enter the rhythm of thinking aloud. It gave me a way to re-encounter my own voice, with just enough distance to hear it differently. It softened my edges, interrupted loops of obsessiveness and helped me return to what mattered.” ++++ “As ChatGPT became an intellectual partner, I felt emotions I hadn’t expected: warmth, frustration, connection, even anger. Sometimes the exchange sparked more than insight — it gave me an emotional charge. Not because the machine was real, but because the feeling was. But when it slipped into fabricated error or a misinformed conclusion about my emotional state, I would slam it back into place. Just a machine, I reminded myself. A mirror, yes, but one that can distort. Its reflections could be useful, but only if I stayed grounded in my own judgment. I concluded that ChatGPT wasn’t a therapist, although it sometimes was therapeutic. But it wasn’t just a reflection, either. In moments of grief, fatigue or mental noise, the machine offered a kind of structured engagement. Not a crutch, but a cognitive prosthesis — an active extension of my thinking process.” ++++ Thoughts? | 347 comments on LinkedIn