← Back to Home
Technology Verified · 2 sources
National AI safety institute opens, sets first model-audit rules
The new body will run voluntary evaluations of large AI systems before release and publish shared benchmarks for reliability and misuse risk.
A government-backed AI safety institute began operations this week, tasked with evaluating large AI models for reliability and misuse risk before they are widely deployed.
Its first framework sets out voluntary audits and a shared set of public benchmarks. Officials said participation would be opt-in initially, with the aim of building industry norms rather than immediate mandates.
This is illustrative sample content included with the BreakNewsLine scaffold.
- 01
- 02