Robust AI Security and Alignment: A Sisyphean Endeavor?
By: Apostol Vassilev
Potential Business Impact:
AI can't be perfectly safe or controlled.
This manuscript establishes information-theoretic limitations for robustness of AI security and alignment by extending Gödel's incompleteness theorem to AI. Knowing these limitations and preparing for the challenges they bring is critically important for the responsible adoption of the AI technology. Practical approaches to dealing with these challenges are provided as well. Broader implications for cognitive reasoning limitations of AI systems are also proven.
Similar Papers
Governable AI: Provable Safety Under Extreme Threat Models
Artificial Intelligence
Keeps super-smart AI from causing disasters.
Never Compromise to Vulnerabilities: A Comprehensive Survey on AI Governance
Cryptography and Security
Makes AI safer and more trustworthy for everyone.
Never Compromise to Vulnerabilities: A Comprehensive Survey on AI Governance
Cryptography and Security
Makes AI safer and more trustworthy for everyone.