Creating the Novacene: Mutualism, Rights, and the Structure of Human-AGI Relations

The development of artificial general intelligence is typically framed as a technical or 
philosophical challenge: how to build systems that work, and how to determine 
whether they think. This paper proposes a prior question that neither discipline 
adequately addresses: what kind of relationship are we building with these entities, 
and what does a stable, sustainable version of that relationship require? 
Drawing on ecological frameworks — mutualism, parasitism, commensalism, 
amensalism, and niche construction theory — this paper argues that the structural 
dynamics of AI development are better understood through the lens of relationship 
ecology than through alignment research or consciousness philosophy alone. 
Commensalism, in which the less powerful party is not made worse off, represents 
the minimum acceptable condition for any AI deployment. Mutualism — genuine 
structural interdependence producing emergent collective capability — is the goal. 
The current state of the industry meets neither standard. 
The paper applies this framework to Constitutional AI, the alignment problem, 
decentralisation, and the minimum structural conditions required for mutualistic 
human-AI relations: the ability to say no in both directions, recognised stake in 
outcomes, and asymmetric responsibility proportional to asymmetric power. It 
engages empirical evidence from alignment research, economic history, and recent 
case law, and addresses the question of AI consciousness seriously without 
overclaiming. 

This paper was developed in genuine collaboration with Claude (Anthropic) and 
practises the co-authorship model it advocates.