Skip to content
Longterm Wiki

Australian AI safety institute announced in 2025, operational early 2026. Budget: A$29.9 million. Three streams: research, engineering, policy/governance. Sits within government as a technical/coordination hub alongside existing regulators and the National AI Centre. Recruiting founding team (General Manager + Head of AI Safety Research and Testing) in late 2025/early 2026.

Related Wiki Pages

Top Related Pages

Approaches

Evals-Based Deployment Gates

Policy

Voluntary AI Safety CommitmentsChina AI Regulatory Framework

Organizations

South Korea AI Safety InstituteCanadian AI Safety InstituteIndiaAI Safety InstituteEU AI OfficeGovAIMETR

Risks

AI Capability Sandbagging

Historical

International AI Safety Summit Series

Concepts

Situational AwarenessCompute Governance

Key Debates

Technical AI Safety ResearchAI Governance and Policy

Other

AI EvaluationsYoshua Bengio