CogniAlign is a survivability-grounded multi-agent deliberation system for AI moral reasoning that reports outperforming GPT-4o on over sixty ethical questions.
Avoid speculative or purely philosophical reasoning unless it is explicitly connected to evolutionary biological empirical evidence
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CY 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
CogniAlign: Survivability-Grounded Multi-Agent Moral Reasoning for Safe and Transparent AI
CogniAlign is a survivability-grounded multi-agent deliberation system for AI moral reasoning that reports outperforming GPT-4o on over sixty ethical questions.