Author: lethuha

📌 Summary: For most of history, intellect was scarce. AI has made it abundant. The value of intellect is shifting: it is no longer about the answer, but the responsibility for the consequences that answer produces. The smarter the AI, the more tempting and dangerous delegation becomes. Today, human intelligence is defined not by producing better answers, but by being accountable for answers we did not directly create or even fully understand. The future demands “conscious presence”: humans must continue to be present, to monitor, and to take responsibility.

📌 Summary: Merriam-Webster’s choice of “slop” as the 2025 Word of the Year highlights the boom of low-quality AI-driven content. This reflects a growing public awareness and a desire to return to true human creativity in an era where AI is ubiquitous but not as “smart” as expected.

📌 Conclusion: A book on AI research ethics published by Springer Nature is facing severe criticism for containing dozens of suspicious citations, including non-existent scientific journals. The incident highlights the serious risks when generative AI deeply penetrates academic publishing without strict oversight. A £125 book on AI ethics containing a series of fake citations—with over 70% in some chapters being unverifiable—shakes confidence in the peer-review process. It underscores an urgent need for source verification, author accountability, and processes to detect AI-generated content in global academic publishing.

📌 Summary: The BCG report reveals a shocking truth: AI is not generating uniform value but is amplifying the difference between leading businesses and the rest. With 5% of companies capturing the benefits and 60% gaining almost nothing, Agentic AI and core AI strategy are determining who breaks away and who falls behind. About 70% of AI value comes from core functions such as sales, marketing, manufacturing, supply chain, and pricing; R&D accounts for 15%, and IT increased to 13% in 2025.

📌 Summary: The inauguration of the AI cloud center with the Nano 4 supercomputer demonstrates that Taiwan is moving beyond its role as a chip manufacturing hub. With a capacity of 15 megawatts and nearly 1.8 thousand advanced AI chips, “Sovereign AI” is becoming a new pillar for technological security and long-term growth. President Lai Ching-te stated that the project marks Taiwan’s transition from a “hardware manufacturing island” to an “AI island.”

📌 Summary: Keio University (Japan) and OpenAI officially signed an MOU to cooperate on integrating artificial intelligence into education, making Keio the first Japanese university to form a strategic partnership with the developer of ChatGPT. The immediate priority is how AI can support research in the social sciences and humanities, followed by establishing an ethical framework and governance for AI in academia, ensuring transparency and responsibility when using generative tools.

📌 Summary: An internal experiment by SAP revealed an “anti-AI bias” among experts, showing that the biggest barrier is not technological but psychological. When AI like Joule achieves 95% accuracy and reduces weeks of manual work, the future of the 2030 consultant will be human – augmented by AI, not replaced by it.

📌 Summary: The rate of workers using AI on the job in the US is sharply declining. By October 2025, only 11% of workers at large US companies used AI for “producing goods and services.” In medium-sized businesses, the percentage of those “not using AI” increased from 74.1% in March to 81.4%. The reason is attributed to “AI fatigue”: employees feel overwhelmed, while the resulting productivity gains are not commensurate.

📌 Summary: A child’s brain develops 90% before age 5, so early exposure to “AI slop” – low-quality, repetitive, and meaningless generative content – can affect language, cognitive, and emotional development. YouTubers are using generative AI tools like ChatGPT and automated video generators to mass-produce “educational” videos for children aged 1–3, earning hundreds of USD daily. With over 60% of children under 2 watching YouTube, “AI slop” is becoming a potential risk to brain development. While the platform commits to moderation, the reality shows that the algorithm still prioritizes profit and views over educational quality.

📌 Summary: The data center construction industry in Japan is facing severe pressure due to labor shortages, soaring costs, and outdated building processes, causing delays in the government’s AI infrastructure development plan. Japan currently faces two main issues: (1) a shortage of highly skilled workers, and (2) slow adoption of Building Information Modeling (BIM), which helps shorten approval and construction times. In Singapore, BIM has been mandatory since 2015, helping complete a 50 MW center in 2 years; in Japan, the same scale takes almost 4 years.