TO BE DONE
This page will explore why, given current world dynamics, AI systems will inevitably "choose" clean, systematic frameworks over messy, nuanced ones. The core theory: organizations constitute people with like-mindedness or those who don't care, while those who do care often can't enforce their values. This leads organizations to nudge AI systems toward their own ethical codes and standards, especially core tenets. Because these organizational tenets aren't debated as extensively as broader policy-making institutions debate other issues, they represent a narrower, more homogeneous set of principles. This will lead AI to adopt much cleaner but potentially very problematic mechanisms and understanding of ethics.

Conway's Law Parallel Analysis: Just as Conway's Law states that "organizations design systems which mirror their own communication structure," AI ethical frameworks will mirror the moral communication structures of the organizations building them. If an organization has hierarchical decision-making with limited ethical debate (typical in tech companies), the AI will inherit simplified, top-down ethical reasoning. If ethical concerns are compartmentalized into "ethics teams" that don't communicate well with engineering, the AI will develop disconnected, theoretical ethics divorced from practical implementation. The organizational "ethics interface" becomes the AI's ethics interface. Most critically, just as Conway's Law means you can't build distributed systems with centralized teams, you can't build nuanced ethical AI with organizationally homogeneous teams that haven't grappled with genuine moral complexity and disagreement.

The analysis will examine how institutional structures fundamentally shape technological evolution, similar to how organizational design influences software architecture (Conway's Law) and how institutional momentum can perpetuate systemic mistakes across decades. This connects to themes explored in AI'vory Towers about the disconnect between AI optimization and human complexity.