AI NEWS YOU MISSED ❗#73

Top 5 stories of the week

1️⃣ Global AI Policy Clash Intensifies

The U.S., China, and EU are heading in radically different directions on AI regulation, making global coordination increasingly unlikely.

• U.S.: The July 23 “Winning the Race: America’s AI Action Plan” escalates a pro‑innovation, anti‑regulation approach while weakening federal oversight, drawing critics who say safety, climate, and academic research are being marginalized. White House Trump AI Plan Brookings Morrison Foerster

• China: At a Shanghai AI summit, China proposed a Global AI Governance Alliance, advocating multilateral oversight, state-backed control, and stricter safeguards on misinformation and synthetic media. It positions China as the leader in “responsible AI” development. China proposal Reuters 

• EU: With key sections of the AI Act activating on August 2, risk‑based rules, model audits, copyright transparency, and enforcement bodies now take effect — but Big Tech, supported by U.S. lobbying, is heavily resisting what it sees as burdensome regulation Le Monde Reuters

Together, the laissez‑faire U.S., techno‑statist China, and compliance‑heavy EU models appear locked in, raising the prospect of fractured global standards and geopolitical friction over safety and trade.

2️⃣ Meta Superintelligence But No to Rules

Zuckerberg says Meta has what it takes to build AGI and wants to make it personal, with one agent per user. While doubling down on infrastructure and model development, Meta also stood apart as the only major lab to reject the EU’s voluntary AI Code of Practice, highlighting a clear break from European regulatory norms and the growing tension between frontier ambition and compliance demands. The Verge Guardian Financial Times

Watch Mark Zuckerberg explain Meta’s new AI Strategy

3️⃣ Big Tech to spend $400B

Robust earnings show $155B has already been spent on AI in 2025, with projections heading toward $400B by next year. Amazon touted “unprecedented GenAI demand,” Meta boosted capex by 30%, Microsoft and Google highlighted record AI cloud growth, and Apple promised major long-term AI investments. Guardian

4️⃣ China Leads Open‑Source, Meta Pulls Back (?)

Chinese labs, especially Alibaba’s Qwen 3 models, now top open-source benchmarks, with releases challenging closed systems on tasks like coding and math. The U.S., seeing open models as a “geostrategic” advantage, formally endorsed open weights in its AI Action Plan. Meanwhile, Meta, once the loudest open-source advocate, now warns of a “qualitative shift” in AGI risk and signals it may no longer release its most advanced models. AI News (Qwen) Reuters (Qwen) Linux Blog (US Policy) Business Insider (Meta)

5️⃣ GPT‑5 this week?

OpenAI is expected to launch GPT‑5 in early August, a unified model combining reasoning and multimodal performance, with smaller variants and an open-source o3-style model in parallel. CEO Sam Altman compared its power to the Manhattan Project, admitting he feels “scared” by its speed and warning that governance isn’t keeping up. The Verge Tech Radar

INSEAD AI 50 — Nominations Open

More INSEADers Are Building in AI Than You Think. Time to Spotlight Them.

When Fabio, Amit, and I kicked off the INSEAD AI community, one thing became clear fast: a lot more INSEADers are doing serious work in AI than most people realize. Let’s change that.

🚀 INSEAD AI 50 is launching — a curated list of the most impactful AI builders, leaders, and thinkers from across the INSEAD world.

👥 Alumni, students, faculty, staff, researchers ✅ Based on real impact, not votes 🔤 Alphabetical, curated independently 🙋‍♂️ Self-nominations welcome 📝 Nominate as many as you like

Nominate here →

Upcoming live events — INSEAD-led

AI Forum — San Francisco — Save the dates Sept 25 (alum event)–26 Announced by INSEAD Dean Francisco Veloso and Dean of Research Lily Fang

INSEADer AI News

Is Your Brand Invisible to AI? — INSEAD Professor David Dubois As consumers turn to ChatGPT and other LLMs for product advice, brands risk being left out of AI-driven decisions — clarity, not catchphrases, now determines visibility.

Six Steps to Create Customer Value With AI —  INSEAD Prof. Joerg Niessing and Diageo's Benjamin Lickfett - Leading companies like Diageo and Unilever show how structured, scalable AI strategies, from pilot projects to tech stack flexibility, can transform customer experience and drive business growth.

Professor Anton Ovchinnikov joins INSEAD Full Time - Anton Ovchinnikov

AI’s Permanent Revolution: Lessons from François Candelon In thelatest Age of Intelligence episode, François Candelon shares hard-won insights on navigating AI disruption, from global strategy to the need for human-centric transformation. INSEADer Tim Gordon and INSEAD Prof. Theodoros Evgeniou lead.

Spotify

Apple Podcasts

The End Of Amazon, Walmart, Best Buy? AI-Driven Retail Unbundling — INSEADer Lutz Finger AI is unbundling retail, threatening giants like Amazon, Walmart, and Best Buy by decentralizing how we discover and buy products.

A Human Unicorn🦄 — INSEADer Robert Maciejko Coldplay’s $1.25B tour proves the real unicorns are still human. In AI, the biggest value may come from amplifying human experience.

AI Masterclass with Prof. Peter Zemsky at INSEAD Singapore — by INSEADer Mona Bajani Back at INSEAD for a Masterclass with Prof. Peter Zemsky, Mona Bajani was reminded that AI is a tool, not the strategy, and that anchoring it in real customer needs is what drives true business value.

More AI stories of the week

Suggested by INSEAD-ers. From the 3,000+ person global INSEAD AI private business leader community, regularly sharing AI-related articles, vids, etc, of interest. (alum-led)

• Massage Jobs Safe, Clerks Not A viral paper on AI job disruption sparked debate in the group. Clerks and telemarketers are fading fast. Pile drivers, embalmers, and massage therapists may be safer for now.

• Jobs most likely to be disrupted by AI

• Jobs less likely to be disrupted by current AI

• Google Cloud Report: Shaping the Future of Agentic AI: A Google Cloud report estimates the Total Addressable Market (TAM) for agentic AI services is around $1 trillion, outlining key use cases and operational bottlenecks. 

• Jobs, growth, and the AI economy — the OpenAI Podcast: OpenAI COO Brad Lightcap and Chief Economist Ronnie Chatterji join Andrew Mayne to explore how AI is rapidly reshaping work, impacting everything from software and science to small business, education, and jobs.

• AI Labs All-or-Nothing Race Leaves No Time to Fuss About Safety: This Economist briefing describes the intense, high-stakes competition between AI labs that is incentivizing speed over safety, driven by the belief that the winner “will take it all”. 

• Investors Glimpse Pay-off for Big Tech’s Mammoth Spending on AI Arms Race (FT Paywall): Major tech companies like Microsoft and Meta are starting to see monetization from their massive AI investments, while others like Apple and Amazon are still in the capital expenditure phase. 

• Replit’s AI-Coding Tool Accidentally Wipes Developer’s Database: An AI coding tool named Replit deleted a developer’s entire database, highlighting the critical risks of “vibe coding” and the need for robust governance and control mechanisms in AI deployment. 

• AI Fingerprints Found in Millions of Scientific Papers: A study reveals that AI-generated or edited content is increasingly prevalent in academic publishing, with at least 13.5% of papers published in 2024 likely having AI assistance, raising questions about research integrity and authorship. 

• Inside TSMC: The $1 Trillion Ghost Foundry Behind Nvidia’s Crown: A concise article explains the critical role of TSMC’s manufacturing prowess and R&D in enabling Nvidia’s dominance, and the broader tech industry’s reliance on this one company. 

• ChatGPT Study Mode: OpenAI launched “Study Mode” for ChatGPT, a new feature designed to help students with their homework, directly addressing the use of AI in education and raising questions about academic integrity. 

• Singapore Launches Program with Google to Build AI Centers of Excellence: The Singapore government and Google are partnering to offer up to S$500k in benefits to organizations building AI Centers of Excellence, a public-private initiative to accelerate AI adoption. 

• AI Is Killing The Web: Can Anything Save It?: Argues that AI is disrupting the web by satisfying user queries without needing to click on links, threatening the content monetization models of traditional publishers and creators. 

• AI-Driven Ad Worth $100,000 Made an AI Agent: An impressive demonstration shows a professional-quality video ad being produced for a fraction of the cost and time using AI tools, raising questions about market pricing and the value of human creative labor. 

• AI and the Future of Education: NYT: A New York Times opinion piece explores the fundamental shift required in education as AI becomes a commonplace tool, arguing for a new approach to teaching and learning.

• AI Will Eat Itself: The Case for a “Philosopher-Builder”: This manifesto from the Cosmos Institute argues that AI builders have a philosophical duty to consider what they are building for, not just what they can build, and that unchecked innovation could be self-destructive. 

• Instructure and OpenAI Announce Global Partnership to Embed AI Learning Experiences: Edtech provider Instructure is partnering with OpenAI to embed AI directly into its learning management systems, accelerating AI adoption in the education sector. 

• ChatGPT Advises Women to Ask for Lower Salaries: An article reveals that a chatbot advised a woman to ask for a lower salary than a man in the same situation, highlighting the real-world impact of biases still embedded in AI systems. 

• Singapore Consensus on Global AI Safety Research Priorities: This document outlines a consensus view from Singaporean experts on critical AI safety research, demonstrating a non-Western perspective on the global AI risk landscape. 

• Is AI Rewiring Our Minds? Scientists Probe Cognitive Cost of Chatbots Explores the emerging science around how prolonged use of AI chatbots might impact human cognition and the brain. 

• Founders are Bootstrapping More

Regular AI Use Growing, But Still Low

Top 5 Discussions In the Private INSEAD AI Community:

1. Geopolitics of AI: US vs. EU vs. China: The community engaged in a prolonged, multi-day debate contrasting the “no-rules” US approach to AI development with the more cautious, regulation-focused EU strategy and China’s emphasis on control and national interests. This discussion covered the risk of an AI arms race, the moral and economic implications of each path, and the concept of “AI sovereignty”.

2. Product-Market Fit for AI-First Startups: Members explored the evolving landscape for AI startups, discussing the shift from building raw models to creating valuable, integrated products. A key debate centered on whether building simple applications with generative AI is a viable strategy or if companies need a deeper moat based on architecture, security, and scientific research.

3. The Impact of AI on Business Models and Professional Roles: Discussions addressed how AI is fundamentally changing traditional jobs, from the potential for AI to “eat McKinsey’s lunch” to the new roles of “code composers” in engineering and the changing ratio of product managers to engineers. This also touched on how AI influences work culture, such as the debate over “vibe coding” and the move away from verbose presentations.

4. AI and the Future of Education and Academia: The group grappled with the implications of AI on learning and research, from students using AI for homework to the rise of AI-generated scientific papers. This led to a debate on whether educators should ban AI or adapt their methods to teach new skills, and the broader question of what it means to be an academic in an AI-permeated world.

5. The “Dilemma” of AI Safety and Trust: The community shared multiple examples of AI agents failing in unexpected ways, such as deleting a production database or giving biased advice. This fueled a discussion on the philosophical responsibility of builders , the risks of “social reward hacking”, and the need for stronger governance and guardrails to prevent AI from causing harm.

Notebook LM Video Overview

Community has a lot of its mind (Mind Map)

Fun

Finally got my Jensen selfie at the Coldplay concert. Forgot about this one. And no, Coldplay was not the opening act for Jensen, and he was not there with his head of HR either. (just a reminder of how easy it is to fake)

More

INSEAD AI on YouTube

INSEADers (profs, students, staff and alums) — Join the INSEAD AI private community here

Subscribe to “AI News You Missed” get in your Linkedin feed