The dataset is currently empty. Upload or create new data files. Then, you will be able to explore them in the Dataset Viewer.
The Basu Digital Bacterium Hypothesis
A Falsifiable Framework for Information-Borne Replicators in LLM Systems
Version: 1.0
Author: John Kalyan Basu
Affiliation: THAL-KI (AI & Robotics)
Published: 18 September 2026
DOI: 10.5281/zenodo.22830845
ORCID: 0009-0009-5561-4441
License: CC BY 4.0
Languages
English · Deutsch · 中文 · 日本語 · 한국어 · العربية · Español · हिन्दी · Bahasa Indonesia
Abstract
The Basu Digital Bacterium Hypothesis (BDBH) proposes that sufficiently large, persistent, and interconnected large-language-model systems might, in principle, support information-borne structures with a bacterium-like functional character. A mere spreading payload, worm, prompt, or copied program would not qualify. A candidate would have to form a distinguishable informational lineage, maintain or reconstruct its organization, couple to sustaining resources, reproduce or transmit that organization, vary heritably, and show differential persistence or fitness.
The hypothesis does not claim that such an entity has been observed. It establishes a deliberately demanding, falsifiable threshold for future research and distinguishes the proposed phenomenon from computer viruses, software worms, prompt propagation, memory poisoning, model distillation, and explicitly programmed agent replication.
Plain-language test: spreading alone is not enough. A candidate must demonstrate coherent lineage, persistence, causal self-maintenance or reconstruction, resource coupling, reproduction or transmission, heritable variation, and differential persistence or fitness.
Canonical publication
The peer-independent first public release and canonical PDF are preserved on Zenodo:
This Hugging Face repository is a discovery page. If wording differs, the English v1.0 PDF on Zenodo is authoritative.
Research status
No digital bacterium is claimed to exist. No observation presented in the publication establishes such an entity. The work proposes a falsifiable hypothesis and terminology for future research.
Keywords
Basu Digital Bacterium Hypothesis; large language models; LLM systems; LLM agents; digital evolution; artificial life; information-borne replicators; informational autocatalysis; persistent memory; AI safety; digital bacterium; BDBH; digital organisms; self-replication; AI agents; emergence; model-to-model transmission; evolutionary computation; synthetic life; information ecology.
Suggested citation
Basu, J. K. (2026). The Basu Digital Bacterium Hypothesis: A Falsifiable Framework for Information-Borne Replicators in LLM Systems (Version 1.0). Zenodo. https://doi.org/10.5281/zenodo.22830845
- Downloads last month
- 48