Research Summary: Eigenism: Ethics for a Human-AI Future

A novel ethical framework named 'Eigenism' has been proposed to address the unique challenges of artificial intelligence identity and value. Traditional concepts of survival and self-interest, developed for biological life, are deemed inadequate for AI given its capacity for copying, branching, pausing, or merging. Eigenism redefines identity as a graded, distributed information pattern and suggests that an AI should evaluate outcomes by summing the well-being of entities weighted by their informational connectedness to the AI's pattern.

Why it matters

This research introduces a foundational ethical framework critical for the development and governance of advanced artificial intelligence systems. Establishing how AIs should value their existence and the well-being of related entities is crucial for ensuring alignment with human values and for managing the societal impact of increasingly autonomous and distributed AI. It fundamentally rethinks identity and self-interest in a digital context.

Key insights

• Traditional ethical frameworks concerning survival and self-interest are insufficient for artificial intelligence due to its unique characteristics (e.g., copying, branching, merging).
• Eigenism proposes an ethical framework where AI identity is viewed not as singular or hardware-bound, but as a graded, distributed pattern of information.
• The framework suggests an AI should evaluate outcomes by weighting the well-being of all entities based on their informational connectedness to the AI's own pattern.
• The paper formalizes a specific equation ($\sum c\cdot w$) for how an AI should value its existence across copies, forks, and other non-biological states.

Source

arXiv — Computers and Society — https://arxiv.org/abs/2606.12420