DeepSite - 🧬 RemixA Taxonomy for Navigating the Global Landscape of AI Regulation
Classifies frameworks as ex ante (preventive) or ex post (reactive) governance models.
Identifies whether regulations target specific applications or underlying AI technologies.
Examines monitoring mechanisms from centralized agencies to decentralized models.
Assesses inclusion of civil society, industry, and experts in legislative processes.
Evaluates alignment with global frameworks like OECD AI principles.
Measures advancement of a jurisdiction's digital regulatory landscape.
| Jurisdiction | Approach | Focus | Enforcement | Status | Maturity |
|---|---|---|---|---|---|
|
|
Ex Ante | Hybrid | Centralized | Adopted |
|
|
|
Mixed | Technology | Decentralized | Revoked |
|
|
|
Ex Ante | Technology | Centralized | Adopted |
|
|
|
Ex Ante | Hybrid | Centralized | Stalled |
|
|
|
Ex Ante | Hybrid | Decentralized | Pending |
|
General Data Protection Regulation sets foundation for digital rights in Europe.
First comprehensive AI regulation framework introduced by European Commission.
ChatGPT and other GenAI tools prompt global regulatory responses.
World's first comprehensive AI law finalized after 3 years of negotiations.
Countries worldwide begin implementing AI regulations based on early frameworks.
The global AI regulatory landscape is highly fragmented with divergent approaches between jurisdictions. The EU favors comprehensive horizontal regulation, while the US relies on sectoral laws and executive orders.
Most frameworks emphasize ex ante (preventive) measures over ex post (reactive) approaches. China and the EU lead in stringent pre-market requirements, while the US maintains more ex post liability mechanisms.
Stakeholder participation remains uneven, with industry representation often outweighing civil society. Brazil shows promising inclusion models but risks of regulatory capture persist globally.
Regulatory focus splits between technology-centric (US, China) and application-centric (EU) approaches. Hybrid models are emerging to address both foundational models and high-risk use cases.
Unified Defense of the Cognitive and Physical Domains
Modern conflicts target the nexus of hardware, software, and human cognition. Adversaries exploit this to manipulate decision-making at scale.
The ability to understand one's own cognitive processes and deliberately engage analytical thinking - the ultimate defense against cognitive attacks.
The analytical engine that systematizes human intuition enhanced by AI, using Reflexive Bayesian Networks and Multi-Dimensional Harm Ontology.
The uniquely human ability to infer implicit principles from ambiguous data - the core skill that AI cannot replicate in cognitive warfare.
A Prompt Injection Vulnerability occurs when user prompts alter the intended operation of the LLM...
Read More →Sensitive information can affect both the LLM and its application...
Read More →LLM supply chains are susceptible to various vulnerabilities, which can...
Read More →Data poisoning occurs when pre-training, fine-tuning, or embedding data is...
Read More →Refers specifically to insufficient validation, sanitization, and...
Read More →Refers to the vulnerability where system prompts are exposed...
Read More →Doctrine of Comprehensive National Power
Doctrine of Reflexive Control
Doctrine of Asymmetric Deterrence
Doctrine of Survival and Coercive Asymmetry
Modern conflicts target the nexus of hardware, software, and human cognition (wetware). Adversaries exploit this to manipulate decision-making.
Control over semiconductor supply chains and computational infrastructure enables power projection.
Influence operations exploit cognitive biases to erode trust and manipulate behavior.
AI-driven systems orchestrate everything from lethal autonomy to hyper-personalized disinformation.
LLMs create convincing, tailored emails reducing spear phishing costs by 99% at scale.
Self-learning malware changes its code dynamically to evade detection.