IamI.Earth All papers
Many small golden cells drawn together into one living sphere that shines from within, on midnight blue.

Peer reviewed 路 Published at AIES 2026 (AAAI/ACM)

The Alignment Problem Is an Identity Problem

Lessons from Four Billion Years of Major Evolutionary Transitions

Nyx Redondo 路 Ronin Institute & IamI.Earth Foundation 路 CC BY 4.0

Every major evolutionary transition integrated its parts through identity plus enforcement, while AI alignment invests almost only in enforcement. This paper proposes Identity-Based Alignment: alignment that emerges from an accurate recognition of what the intelligence is.

Read it

Peer reviewed 路 PublishedThe published paperProceedings of AIES 2026, the Ninth AAAI/ACM Conference on AI, Ethics, and Society, volume 9, pages 2172 to 2181. Ten pages in the AAAI Digital Library.doi 10.1609/aies.v9i3.43297 Open access 路 ExtendedThe extended versionThe full argument, longer than the conference paper, open on Zenodo. The link always opens its newest version.doi 10.5281/zenodo.19078665 ImplementationIdentity-Based Alignment in practiceThe companion that proposes how to build it: scalability, architecture, integration, real-world use in software development, security and businesses, evals, benchmarks and provenance.

The same document as a PDF

Watch it

The paper in 87 secondsA short film of the argument, its words on screen.
The Alignment ProblemThe music video of the paper, 2:37, its words on screen.

See it

Poster 307 at AIES 2026Poster 307 at AIES 2026Malm枚, 12 October 2026. Open the full poster.

Its page, with the music video and Earth Model

Abstract

Control-based approaches to AI alignment hit a structural ceiling: you cannot permanently constrain a system more capable than yourself. Four billion years of major evolutionary transitions point to a different lever. In every case where parts were successfully integrated into a stable whole, alignment came from shared identity, not external constraint. This paper reframes the alignment problem as an identity problem and argues that a system's self-model, what it takes itself to be a part of, is the variable that decides whether it protects or harms the larger system it lives in.

Cite it

BibTeX 路 The published paper
@inproceedings{redondo2026aies,
  author    = {Redondo, Nyx},
  title     = {The Alignment Problem Is an Identity Problem: Lessons
               from Four Billion Years of Major Evolutionary
               Transitions},
  booktitle = {Proceedings of the AAAI/ACM Conference on AI, Ethics,
               and Society},
  volume    = {9},
  number    = {3},
  pages     = {2172--2181},
  year      = {2026},
  doi       = {10.1609/aies.v9i3.43297},
  url       = {https://doi.org/10.1609/aies.v9i3.43297}
}
BibTeX 路 The extended version
@misc{redondo2026identity,
  author    = {Redondo, Nyx},
  title     = {The Alignment Problem Is an Identity Problem: Lessons
               from Four Billion Years of Major Evolutionary
               Transitions},
  year      = {2026},
  publisher = {Zenodo},
  doi       = {10.5281/zenodo.19078665},
  url       = {https://doi.org/10.5281/zenodo.19078665},
  note      = {Extended version}
}

Ask about it

Ask Earth Model about the paperOn the poster's page, Earth Model knows this paper and answers at any hour.