https://seunghoonchoi.com/tags/ai-safety/Recent content in AI Safety on Seunghoon ChoiHugoen-USThu, 25 Jun 2026 23:40:00 +0900https://seunghoonchoi.com/column/kim-joungho-ai-constitution/Thu, 25 Jun 2026 23:40:00 +0900https://seunghoonchoi.com/column/kim-joungho-ai-constitution/<p><img src="https://seunghoonchoi.com/images/col-kim-joungho-ai-constitution.jpg" alt="Server racks in a server room, suggesting the electricity and control behind AI systems"></p> <p class="inline-image-caption">The constitutional format does not leave AI discussions in the form of abstract slogans, but divides them into responsibilities and powers for each article.</p> <p>The text below is a reconstruction of the core remarks made by KAIST School of Electrical Engineering Professor Joungho Kim in his 2025 Kim Dae Jung Peace Forum special lecture, &ldquo;Effective Development of AI and Peace Promotion,&rdquo; rewritten in the form of constitutional articles.</p>https://seunghoonchoi.com/column/why-would-agi-obey/Sun, 21 Jun 2026 19:30:00 +0900https://seunghoonchoi.com/column/why-would-agi-obey/<p><img src="https://seunghoonchoi.com/images/col-why-would-agi-obey.jpg" alt="Chess pieces set out on a board in a stone-walled room, under light from a window"></p> <p class="inline-image-caption">What is scarier in the AGI debate is not the hostility, but the possibility that the system may continue to operate even after humans are no longer needed.</p> <p>One question wouldn&rsquo;t leave my head. An AGI overwhelmingly smarter and stronger than humans: why on earth would it obey us? After digging for a while, the answer I came back with was a little deflating. There&rsquo;s no reason it would. And the more I chewed on it, the more that empty feeling started to seem obvious.</p>