Power is never static. It flows, mutates, and embeds itself within the structures we deem natural. Today, it flows through large language models (LLMs)—those engines of meaning that can organize knowledge and could soon dictate how humans interpret the world. WorldView is an investigation into the unseen mechanisms of power embedded within LLMs.
We acknowledge that even open-source LLMs, despite their accessibility, remain inscrutable in how they generate narratives and produce truth claims. Their lack of explainability is not a technical flaw but a feature of their epistemic dominance, allowing them to shape cultural, geopolitical, and ideological paradigms without question. Our project deconstructs this opacity by mapping how AI narratives emerge, spread, and reinforce global structures of influence.
WorldView is essential for everyone involved in funding, developing, regulating, and using LLMs. With a particular focus on international peace, WorldView reveals how AI-crafted narratives become tools of persuasion, how they marginalize alternative perspectives, and how they consolidate power within certain ideological frameworks. In a world where knowledge itself can be manufactured, WorldView equips strategic thinkers with the power to see the invisible contours of AI’s influence and find ways to use it for world peace. This is definitely not the end of human civilization—but the beginning of its most critical test.
