AI Research Memo

Judge AI use by proportion, not yes/no Require disclosure and provenance to prove human lead Apply thresholds (≤20%, 20–50%, >50%) to grade and govern Sixty-two percent of people say they would like their favorite artwork

The real risk isn’t the LLM’s words but the agent’s actions with your credentials Malicious images, pages, or files can hijack agents and trigger privileged workflows Treat agents as superusers: least privilege, gated tools, full logs, and human checks

AI lowers entry barriers, raises mastery standards Novices gain most; experts move to oversight and design Education must deliver operator training and governance mastery A quiet result from a very loud technology dese

AI use is ubiquitous; current assessments reward fluency over thinking Grade process, add brief vivas, and require transparent AI-use disclosure Train teachers, ensure equity, and track outcomes to make AI a partner Eighty-

AI accelerates information cascades, turning rumors into rapid bank runs Stability now hinges on dampening synchronized behavior, not just capital buffers Build rumor-aware stress tests, fast disclosures, and drill-based curricula

U.S.

AI translation reshapes the labor market, eroding low-skill roles while rewarding domain expertise Education must shift toward “language plus” skills—pairing translation with data, law, or health Policy should teach students to work with machines, not against them

AI and robotics remain narrow tools, excelling only in tightly defined tasks Human versatility—handling exceptions, combining roles, and adapting to context—remains the decisive advantage Education policy must prioritize training for this versatility, turning automation into complement rather than substitute

Search behaves like reinforcement learning, rewarding confirmation Narrow queries and clicks shrink exposure at scale Break the loop with IV-style ranking and teach students to triangulate queries

Student well-being is falling fast AI chatbots are spreading quickly Without safeguards, risks will escalate

AI’s IMO gold isn’t AGI Deploy it as an instrumented calculator Require refusal metrics and proof logs In J

LLMs are not conscious, only probabilistic parrota They often mislead through errors, biases, and manipulations Education must use them as tools, never as advisors

I am in my early 40s and work at an office near Magok Naru Station and I live near Haengsin Station in Goyang City. I used to commute by company shuttle, but recently I've taken up cycling as a hobby and now commute by bike. The biggest reason I got into cycling was because of the positive image I had of Seoul's public bicycle program, Ddareungyi.

While founding a university (SIAI), I encountered a surprising reality—university rankings, like any evaluative system, are shaped by more than just academic performance. Factors such as institutional branding, media visibility, and methodological choices play a role in shaping how institutions are perceived and ranked. This has led to ongoing debates about how rankings should be structured and whether certain metrics introduce unintended biases.

I have spent years in AI and data science, believing that structured models and quantitative analysis were the future. That perspective changed the moment I became a target of an orchestrated misinformation campaign—one that wasn’t random but designed to destroy my credibility, my institution’s reputation, and my work.

80% of Korean students failed at SIAI not due to lack of intelligence but due to deep-rooted cultural conditioning that discourages independent thought and risk-taking The Confucian, exam-based education system promotes rote memorization over problem-solving, making students struggle in an environment that requires deep, abstract thinking Korea’s broader economic and corporate structure reinforces a ‘safe thinking’ mindset, making it unlikely that Western-style innovation will thrive here without significant systemic change Before going into details, pl

Not due to math knowledge, but due to difficulty applying knowledge in real-world scenarios accustomed to structured learning, struggle more with open-ended, problem-first approaches compared to those trained in Western-style superficial engagement, reliance on structured guidance, avoidance of ambiguity, and resistance to open-ended problem-solving Failed in abstraction (encoding) and application (decoding) Since 2021, the Swiss Institute of Artificial Intelligence (SIAI) has refined its approach to teaching AI and data science (DS), learning valuable

AI Bootcamps provide emotional satisfaction but no real AI knowledge. SIAI’s AI MBA (Business & Tech Tracks) offers real AI project exposure and strategic thinking. Basic software engineers will be obsolete by 2035, replaced by AI and offshore talent After launching AI MBA's business track, we sometimes have questions about the value of the track. Most people, particularly, engineers think that's just a waste of time. Some of them even claim that AI Bootcamp is the better option, as it costs less money.

Over the past few years, there has been a growing trend of STEM MBAs—business programs that integrate basic AI, analytics, and coding to appeal to professionals interested in tech-driven industries. While these programs may sound promising, in reality, most STEM MBAs provide little more than bootcamp-level technical training, leaving graduates with surface-level AI knowledge and little ability to differentiate real AI innovation from hype.