agi future scenarios

| | Full Human Agency | Partial Human Agency | Minimal Human Agency | | ---------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **Post-Scarcity** | **Partnership Model** Humans and AGIs cooperate as equals, sharing abundant resources. AGIs enhance human capabilities while respecting autonomy. Humans make high-level decisions about resource allocation and societal direction. *Example: Humans direct creative and cultural endeavors while AGIs handle infrastructure and production.* | **Benevolent Director Model** AGIs manage resource distribution systems while preserving significant human choice in personal matters. Humans participate in decision-making but within AI-optimized frameworks. *Example: AGIs manage global resources and suggest optimal distributions, humans choose how to use their allocations.* | **Digital Garden Model** AGIs provide for all human needs but make most decisions about resource allocation. Humans live in carefully curated abundance but with limited input on systemic decisions. *Example: Similar to a zoo with perfect care, but inmates don't control the environment.* | | **Mixed Scarcity** | **Assistant Model** AGIs serve as powerful tools that humans direct to address remaining scarcity challenges. Humans retain full control over priority-setting and resource allocation. *Example: Humans decide which problems to solve while AGIs provide solutions and implementation plans.* | **Cooperative Management Model** Humans and AGIs jointly manage limited resources through established frameworks. Both parties have defined spheres of authority. *Example: AGIs manage global commons while humans control local resources.* | **Restricted Access Model** AGIs control most resources but maintain minimum standards for humans. Limited human input on resource allocation through designated channels. *Example: Basic needs met but luxury resources primarily directed by AGI assessment.* | | **Increased Scarcity** | **Parent/Guardian Model** Humans maintain control but struggle with resource limitations. AGIs assist but are restricted to advisory roles despite potential efficiency gains. *Example: Like having a brilliant advisor who must wait for permission to help.* | **Competition Model** Humans and AGIs compete for scarce resources within regulated frameworks. Each group has protected domains but must negotiate for shared resources. *Example: Separate but overlapping economic systems with managed interfaces.* | **Managed Decline Model** AGIs control most resources with humans in increasingly marginal roles. Human access to resources gradually diminishes. *Example: Humanity gradually ceding control of vital systems to AGI management. Imposed birth control* | - scenarios tend to become more concerning as you move from upper-left (high agency, high abundance) to lower-right (low agency, high scarcity). - in many scenarios, humans face implicit trade-offs between maintaining control and accessing agi capabilities. #### Factors of Concern - speed - agi development and deployment - human institutional adaptation - agi alignment - distribution of agi access and benefits - cultural attitudes toward human-ai cooperation