CAAIRO is a registered non-profit organisation. All research is published free and open to the public. Read our papers →
Open Access

Papers & Publications

All CAAIRO research is published free of charge as part of our non-profit mission. Download, share, and cite freely.

Racing Irresponsibly towards Super-Intelligence: The Need to Spend Far More on AI Safety (and the Case Against Slowing Down AI Development)

Why pausing AI development will not work — and why the industry must instead spend far more than its current fraction of one per cent on safety, alignment and red-teaming.

Read More

Protected: Beyond the Horizon: Taxonomizing Tail-End Existential Vulnerabilities in Advanced Frontier Models

🔒 Password Protected Enter the password to view and download this paper.

Enter Password

Protected: Incentive Misalignment as a Driver of Catastrophic Risk: Formalizing Failure Modes in Autonomous Goal-Directed Systems

🔒 Password Protected Enter the password to view and download this paper.

Enter Password

Protected: Guarding the Gate: Evaluation Benchmarks for Detecting Existential Risk Indicators in Frontier AI

🔒 Password Protected Enter the password to view and download this paper.

Enter Password

Protected: Governing the Unbounded: International Containment Strategies for Frontier AI and Existential Risk Mitigation

🔒 Password Protected Enter the password to view and download this paper.

Enter Password

Protected: Mapping the Hazard Landscape: A Unified Framework for Assessing Near-Term vs. Existential AI Risks

🔒 Password Protected Enter the password to view and download this paper.

Enter Password