Organization · Nonprofit

Singapore AI Safety Hub (SASH)

An independent nonprofit building Singapore and Asia's AI safety field; with international partners, it develops a confidential network logger for inference verification.

www.aisafety.sg · also called SASH

SASH describes itself as "an independent nonprofit building Singapore and Asia's AI safety field". Its verification work is an international collaboration on inference verification:

  • Collaboration. SASH reports launching, with the Oxford Martin AI Governance Initiative and a team of researchers, an international collaboration to build tools that make AI verifiable across borders, and names the Future of Life Institute as another early partner 1.
  • Confidential network logger. The first project asks whether confidential network loggers can help distinguish inference from training 1. SASH describes loggers that copy the traffic leaving a data centre; in its planned next prototype, a random sample of requests goes to a separate cluster that re-runs them and raises an alarm on a mismatch 1. See SASH confidential network logger, Network taps and certifiers and Sampled inference recomputation.
  • Code. The first prototype's code is public under the MIT licence 2.
  • Field listing. The AI Futures Project's verification page lists FLI and SASH as building the logger prototype 3.

Implementations

  • An open-source prototype that routes a facility's inference traffic through a logger and re-runs requests on a separate cluster to check it serves inference.

Mechanisms

Mechanisms this organization has designed, built, evaluated or supplied.

  • Devices on a cluster's network links that copy and hash all traffic, so a verifier can later check sampled records against declared work.

Publications

Sources this organization authored or published.

Sources

  1. CSingapore AI Safety Hub (SASH) (2026). Internationalising AI Verification. SASH blog. Source recordSupports: international collaboration and partners; confidential network logger design and planned prototype
  2. BSingapore AI Safety Hub (SASH) (2026). inference-verification: Inference Verification Prototype. GitHub. Source recordSupports: prototype code (MIT licence)
  3. CAI Futures Project (2026). Get Involved in Verification. AI 2040. Source recordSupports: listed with FLI as building the logger prototype