引導式測驗 · 中等 等級
世界各地的人工智慧安全研究所測驗
測試英國人工智慧安全研究所和美國 CAISI,部署前測試的工作原理以及這些機構的法律權力限制。
問題 1 的 8
What was the UK AI Safety Institute renamed in February 2025?
預覽每個問題8 問題
當您參加測驗時,就會出現答案和解釋。
- What was the UK AI Safety Institute renamed in February 2025?
- AI Standards Office
- AI Security Institute
- Frontier AI Regulator
- National AI Agency
- The US AI Safety Institute was reorganized into which body in 2025?
- Center for AI Standards and Innovation (CAISI)
- Federal AI Commission
- Office of AI Safety Enforcement
- National AI Research Resource
- Where did the International Network of AI Safety Institutes hold its first meeting?
- London
- Seoul
- San Francisco
- 巴黎
- What legal power do the UK and US institutes generally have over model releases?
- They can issue fines for any unsafe model
- They must certify every model before sale
- They can seize model weights
- They have no general statutory power to block releases and rely on voluntary access
- What is Inspect?
- An open-source AI evaluation framework released by the UK institute
- A US law requiring model audits
- An EU list of banned AI practices
- A commercial chatbot
- Which organization has actual enforcement powers over general-purpose AI models under the AI Act?
- UK AI Security Institute
- US CAISI
- EU AI Office
- Japan AI Safety Institute
- Why does strong elicitation matter in pre-deployment testing?
- It shortens the model's training time
- It reduces API costs
- It is needed to publish results
- Without tools, scaffolding and many attempts, tests may underestimate what a model can do
- Which event did the UK institute's creation coincide with?
- The Paris AI Action Summit
- The Bletchley Park AI Safety Summit
- The G7 Hiroshima Summit
- The UN AI Advisory Body report