Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
September 14, 2026Open Access

Emotions and a Functional Test for Intelligence in Artificial Systems

View Full Paper
Ask AI
Bookmark
Share

Authors

JTJeroen Terpstra

Discussion

Loading...

Member takes

Overview

Theoretical analysis outlines three functional criteria for artificial emotion, suggesting genuine artificial intelligence requires coordinated reorganization, plasticity, and internal stakes.

Key Points

  • To establish a rigorous, falsifiable, and operational functional test for the presence of emotions and genuine intelligence in artificial computational systems.
  • Critiqued existing functional frameworks of artificial emotion regarding fixed computational architecture, substrate independence, and persistent internal states.
  • Developed a layered functional model of cognition to derive operational requirements for emotional processing in machine learning networks.
  • Identified three jointly necessary conditions for artificial emotion and intelligence: processing reorganization, learning-signal plasticity, and vulnerable internal stakes.
  • Determined that current large language models do not exhibit emotions because their weights remain frozen post-training and they lack persistent internal states requiring protection.

Cite This Study

Jeroen Terpstra (2026) studied this question.

synapsesocial.com/papers/6aa7b41e0926e14a848b3adbhttps://doi.org/10.5281/zenodo.22737021
View Full Paper
Ask AI
Bookmark
Share