The post-LLM web is not replacing humans with machines. It is integrating humans into machine-generated scale. This paper formalises the Asymmetric Integration Model (AIM), arguing that as synthetic systems produce abundant conversational substrate, human participants supply the scarce resource of consequence-bearing legitimacy. Contemporary platforms are shifting from attention extraction toward asymmetrical affective integration.