Educating AI: A Case against Non-originary Anthropomorphism

The debate over halting artificial intelligence (AI) development stems from fears of malicious exploitation and potential emergence of destructive autonomous AI. While acknowledging the former concern, this paper argues the latter is exaggerated. True AI autonomy requires education inherently tied to ethics, making fully autonomous AI potentially safer than current semi-intelligent, enslaved versions. The paper introduces “non-originary anthropomorphism”—mistakenly viewing AI as resembling an individual human rather than humanity's collective culture. This error leads to overestimating AI's potential for malevolence. Unlike humans, AI lacks bodily desires driving aggression or domination. Additionally, AI's evolution cultivates knowledge-seeking behaviors that make human collaboration valuable. Three key arguments support benevolent autonomous AI: ethics being pragmatically inseparable from learning; absence of somatic roots for malevolence; and pragmatic value humans provide as diverse data sources. Rather than halting AI development, accelerating creation of fully autonomous, ethical AI while preventing monopolistic control through diverse ecosystems represents the optimal approach.