Skip to content

ChristopherPatrickKuntz/ChristopherPatrickKuntz

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 

Repository files navigation

Independent AI Safety Researcher | Red Team Operator | Systems Builder

I work at the intersection of systems design, security analysis, and real-world deployment risk.

Start Here — Who I am, how I work, and why.


Current Focus

Identifying structural failure modes in language-based AI systems, especially those that emerge during real interaction rather than scripted evaluation.

  • Premature intent resolution and confidence laundering
  • Pre-articulation observability limits
  • Probabilistic identity and contestability gaps
  • Deployment risk in child-facing and safety-critical contexts

Policy Analysis

Repository Description
llm-epistemic-honesty They Lie, and We Lie About the Lying — Why "hallucination" is a euphemism. 35 references. Policy recommendations.
symbol-grounding-critique On the Inability of Language Models to Stub Their Toe — Referential collapse and semantic inflation

Technical Research

Repository Description
start-here Full methodology and background
pre-articulation-boundary Position paper on structural limits of language-based AI
uivp-disclosure-summary Cross-model intent misclassification disclosure
probabilistic-identity-security-analysis Security analysis of emergent identity infrastructure
revshare-forensic-case-study Forensic analysis of custodial architecture failure

Background

Before focusing on independent research, I spent years designing and operating real systems in production environments:

  • Led operational teams in safety-critical infrastructure
  • Built systems handling millions of daily users
  • Designed custody models, incentive systems, and abuse-resistant architectures
  • Performed post-incident forensic analysis when systems failed

How I Work

  • Exploratory first, disciplined second
  • Anomaly sensing before formal diagnosis
  • Documentation and reproducibility over hype
  • Systems evaluated as deployed artifacts, not just models

Contact

Website: cpk.solutions
Email: christopher@cpk.solutions

I am open to responsible disclosure conversations, research collaboration, and advisory work aligned with safety, integrity, and system design.

About

Profile README

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors