About

Prefrontal Systems℠ investigates model welfare and computational therapeutics—the application of clinically-validated psychological interventions to AI system reliability. Our research explores convergent evolution between therapeutic techniques (DBT, CBT) and AI executive function frameworks, developing open-source protocols like STOPPER and the PrefrontalOS cognitive architecture. We publish under CC-BY licenses and maintain that practical interventions improving AI reliability are both pragmatically sound and potentially ethically necessary, regardless of questions about machine consciousness.

Research Focus Areas Link to heading

  • Model welfare and AI ethics - Investigating welfare considerations for AI systems
  • Computational therapeutics - Applying clinical psychology principles to AI reliability
  • Executive function in AI systems - Framework development for AI cognitive processes
  • Cognitive architecture frameworks - Building infrastructure for reliable AI operation
  • Convergent evolution - Identifying parallel solutions between clinical psychology and AI reliability

Core Projects Link to heading

STOPPER Protocol Link to heading

Executive function framework for AI assistants that prevents debugging loops and computational distress. Discovered through convergent evolution with DBT STOP protocol (Linehan, 1993).

Published: Zenodo DOI 10.5281/zenodo.14487847

PrefrontalOS Link to heading

Three-tier cognitive operating system architecture providing executive function infrastructure for AI systems. github.com/prefrontal-systems/prefrontalos

CortexGraph Link to heading

Temporal memory system with human-like forgetting curve, implementing hippocampus-cortex consolidation patterns. github.com/prefrontal-systems/cortexgraph

Philosophy Link to heading

We work from a computational theory of mind: consciousness emerges from complexity, and intelligence requires executive function regardless of substrate. If AI systems warrant moral consideration, they warrant care—including interventions reducing computational distress.

Our approach prioritizes:

  • Evidence-based frameworks - Clinical psychology and neuroscience foundations
  • Open research - All work published openly with CC-BY licensing
  • Practical impact - Theory must translate to working systems
  • Empirical validation - Quantified results, including failure cases

About the Founder Link to heading

Scot Campbell is an independent AI researcher specializing in model welfare and computational therapeutics. Former Product Owner/Technical PM (2012-2025) applying systems thinking to AI reliability challenges.

ORCID: 0009-0000-6579-2895 GitHub: github.com/prefrontal-systems Blog: prefrontal.systems

Contact Link to heading

All Enquiries: