Today's large language models (LLMs) are produced by data from the public, so they can't ethically be owned by any one individual. Socialism is the only fair way to govern this technology.