Published July 7, 2026
| Version v1
Preprint
Open
Engineering Explicit Consciousness: A Safety Primitive for Observable Metacognition in LLM Agents
Authors/Creators
-
1.
Università Campus Bio-Medico
-
2.
University of Chieti-Pescara
Description
This preprint proposes a safety primitive for observable metacognition in LLM agents, introducing a framework to make machine self-awareness detectable and verifiable. The work challenges current black-box approaches to AI consciousness and offers a concrete engineering methodology.
Files
paper260707.pdf
Files
(651.1 kB)
| Name | Size | Download all |
|---|---|---|
|
md5:f95eb22aaa050e1738399de9e6ca62a8
|
651.1 kB | Preview Download |