IAIGCRM Cypress emblem
IAIGCRMINSTITUTE FOR AI GOVERNANCE, CYBERSECURITY & RISK MANAGEMENT
Research & Publications / Internal editorial preview

Existential Risk
& AI Safety

Independent research. Global perspective. A deeper understanding of humanity’s future.

An interdisciplinary research initiative dedicated to studying and understanding existential, global catastrophic, and systemic risks associated with advanced artificial intelligence and related technologies. Our research examines emerging threats, potential consequences, interconnected risk domains, and the evolving body of scientific and scholarly evidence. Through structured taxonomies, comparative analysis, and critical examination of governance and safety frameworks, we seek to advance understanding, encourage informed discourse, and contribute to the broader international research community.

Explore the Framework →About This Research
Website v2.3.1 · Research preview
01

Existential Risk Taxonomy

Fourteen ER domains, permanent identifiers, mechanisms and cross-domain references.

Explore taxonomy →
02

Risk Assessment & Analysis

Scenario modeling, cascading effects, uncertainty and escalation pathways.

Explore assessment →
03

Research Sources & Attribution

Scholarly bibliography, provenance verification and taxonomy mapping.

Explore sources →
04

Governance, Safety & Societal Responses

Examination of governance frameworks, proposed safety approaches, institutional responses, and international cooperation.

Explore governance →
Research caution. Taxonomy entries identify candidate risk mechanisms, not established existential threats. Source associations are preliminary, not endorsements by cited authors. The v0.2 research baseline remains authorized internally; v0.3 is under editorial review. No public publication has been authorized.
14Risk domains
92Candidate mechanisms
v0.3Taxonomy working draft

Comprehensive ER Taxonomy

Expand any domain to view its permanent subdomain identifiers and associated research sources.

Featured Research Scholars

Selected contributors to AI safety and existential-risk scholarship; inclusion does not imply endorsement of IAIGCRM.

Nick Bostrom

Superintelligence (2014) · Long-term AI risk

Stuart Russell

Human Compatible (2019) · AI alignment and control

Roman V. Yampolskiy

AI: Unexplainable, Unpredictable, Uncontrollable (2024) · AI control

Yoshua Bengio

Managing Extreme AI Risks Amid Rapid Progress (2024, coauthor) · AI safety

Geoffrey Hinton

Managing Extreme AI Risks Amid Rapid Progress (2024, coauthor) · Advanced AI risk

Paul Scharre

Army of None (2018) · Autonomous weapons

Research Sources & Attribution

Expanded curated research register. Browse by author, domain or recent publications. All entries require final editorial verification.

02 / Risk Assessment & Analysis — Methodology

Specify threat scenarios, initiating conditions, causal and escalation pathways, systemic impacts, assumptions, uncertainty, confidence and residual risk. Distinguish evidence-based analysis from speculative mechanisms.

04 / Governance, Safety & Societal Responses — Methodology

Examine proposed preventive, detective and responsive safeguards; compare governance arrangements, accountability, oversight and international cooperation. Distinguish demonstrated effectiveness from proposals, unresolved questions and residual uncertainty.

Methodology & Publication Control

Research records distinguish empirical evidence, historical analogy, modeled analysis, theoretical argument and speculation. Consequence tiers (systemic, global catastrophic and existential) are not probability estimates. Capability maturity and uncertainty must be documented independently.

Identifier policy: ER-13 and ER-14 appear immediately after ER-03 in thematic display order. Permanent identifiers do not change. Cross-domain overlaps are retained pending editorial resolution.

Version control: Website v2.3.1 landscape refinement; taxonomy v0.3 draft; source register v0.2 candidate. Publication requires separate authorization.