Author: lethuphuong

📌 The new draft regulation shows that China wants to strictly control AI with emotional capabilities, viewing psychological risks and addiction as critical issues. By requiring user emotion assessment, overuse warnings, and active intervention, Beijing is laying an early legal foundation for “human-like” AI. This could shape how AI businesses design products in China and influence global AI governance standards in the future.

📌 Electric Twin develops AI polling technology, using AI-controlled fake “voters” or “customers” trained on real-world data and emotions to provide rapid feedback. The company claims 91% accuracy, comparable to traditional polling methods. Raising nearly £7 million from Atomico reflects strong investor confidence in the AI-generated “synthetic person” model. However, this technology also raises questions about ethics, representation, and the risks of policy decisions based on data not derived from real humans.

📌 Once again, Deloitte – one of the Big Four – is being criticized for errors related to the use of AI in a consulting report for the government. The errors included inaccurate descriptions and misidentification of hospitals as well as healthcare facilities. After being detected, Deloitte stated it “stands fully behind the recommendations” made in the report, despite the errors caused by generative AI, which were primarily in the citations section. The incident raises concerns about the reliability of AI-assisted consulting reports in the public sector.

📌 PayPal’s Global AI Head believes that humanity has left the Information Age to enter the “Intelligence Age,” where AI not only stores and retrieves data but also generates new data itself. In the Information Age, value came from data access; in the Intelligence Age, the focus is on generation, reasoning, and moving towards partial automation of work.

📌 OpenAI CEO Sam Altman believes that the next major breakthrough on the path to superhuman artificial intelligence will come from memory, not reasoning capabilities. With a vision of AI being able to remember a user’s entire life, OpenAI aims to create an assistant that truly “participates in human life” starting in 2026. In the context of ChatGPT gradually losing market share to Gemini and new competitors, AI memory could become the strategic weapon deciding the leader in the next phase of generative AI.

📌 Biren Technology, a Shanghai-based AI chip manufacturer, plans an IPO in Hong Kong aiming to raise up to 4.85 billion Hong Kong dollars (≈623 million USD). Biren is one of China’s “Four Little Dragons” of GPUs, alongside Moore Threads and MetaX. With previous deals surging 400–800% and total Hong Kong IPOs reaching 259.4 billion HKD in 2025, investors are betting big on chip autonomy and AI. Despite not yet being profitable, AI companies are seen as the main drivers for the new IPO cycle in Asia.

📌 The 10% layoff wave at McKinsey shows that the elite analysis-based consulting model is reaching its limit in the AI era. AI not only reduces information asymmetry but also “flattens” analytical and recommendation capabilities – the core value foundation of strategic consulting. The value focus is shifting from strategy to technology-based execution, where strategy and execution occur simultaneously and continuously.

📌 The phenomenon of “cognitive offloading” – outsourcing thinking to tools like AI – is becoming widespread and could erode core thinking capabilities. When given a choice between thinking for themselves or using AI, humans often choose AI, even when they are fully capable of solving the problem themselves. The cause is metacognitive: humans underestimate their own abilities, leading to a habit of avoiding mental effort. Smart people face a paradox: they leverage AI best but also risk losing the training process that created that intelligence.

📌 The phenomenon of “cognitive offloading” – outsourcing thinking to tools like AI – is becoming widespread and could erode core thinking capabilities. When given a choice between thinking for themselves or using AI, humans often choose AI, even when they are fully capable of solving the problem themselves. The cause is metacognitive: humans underestimate their own abilities, leading to a habit of avoiding mental effort. Smart people face a paradox: they leverage AI best but also risk losing the training process that created that intelligence.

📌 AI is fundamentally changing the working models of top consulting firms like McKinsey, BCG, PwC, EY, Deloitte, and Accenture. The focus is shifting from short-term consulting projects and slide decks to multi-year AI transformation programs, requiring capabilities in technology construction and operation. Engineers and tech experts are increasing rapidly. Competitive advantage now lies in hybrid personnel: understanding business, knowing how to deploy AI, communicating well, and learning extremely fast. The race is not just about hiring more engineers, but retraining hundreds of thousands of people to consult in the AI era.