Develops a non-operational failure-mode framework for how advanced AI could cause extinction or permanent disempowerment via misalignment, human misuse, organizational failure and competitive deployment, without requiring intent.
Argues that proposed AI sanctuaries for rogue agents would create selection effects and resource dynamics that undermine their goals, and that the study-and-release plans are constrained.
A beginner-oriented FAQ defending AI x-risk arguments: misalignment, instrumental convergence, shutdown resistance, misuse, and why governance action is urgent.
A personal backchaining essay on what positive AGI futures require, arguing for pluralistic, individually controlled and aligned AI plus a hardened, defensive-acceleration society against catastrophic risks.
Argues that AI accelerating science raises near-term biological extinction risk, that aligned-ASI arguments for rushing development fail, and that specifying ASI behaviour remains unsolved.
Proposes Stackelberg Security Games as a framework for AI oversight, covering training-time auditing against poisoning, constrained pre-deployment evaluation, and robust multi-model deployment.
Argues that in multipolar ASI offense/defense scenarios, defensive AI aligned with humans who fail to colonize space could let offense win, and proposes a minimum viable defense theorem.
A broad framework mapping AI risks from current harms to existential threats, organised into misuse, misalignment and systemic categories plus risk amplifiers, linking observed behaviours like reward hacking and scheming to future catastrophe.
Analyses whether intent-aligned defensive AI can counter takeover-level adversarial AI, proposing threat models, a three-pillar defense framework, and legal constraints, and concludes the offense-defense balance remains uncertain.
Conceptual paper that defines and separates AI safety from AI security, outlines their research boundaries and interdependencies, and argues the clarification aids research, collaboration and policy.
Nielsen essay arguing that the core ASI existential danger is the raw power and lowered barriers to dangerous technologies, not rogue AI, and that alignment work may speed catastrophic capabilities.
Uses formal language theory and data on the LLM ecosystem to show that LLMs behind adversarial outputs are often unidentifiable from text alone, and that candidate origins grow exponentially, making attribution infeasible.
An overview that organizes catastrophic AI risks into malicious use, AI race, organizational risks, and rogue AIs, describing hazards and proposing mitigation directions for each.