{
    "@id": "https://newmindmuseum.com/topics/ai-safety-testing/#entity",
    "id": "tema/ai-safety-testing",
    "type": "DefinedTerm",
    "name": "AI Safety Testing",
    "description": "How labs and independent groups test models for risky capabilities before and after release.",
    "url": "https://newmindmuseum.com/topics/ai-safety-testing/",
    "sameAs": [],
    "wikidata": "",
    "stories": [
        {
            "title": "Tests Show AI Flagged Iranian School as Military Target",
            "url": "https://newmindmuseum.com/safety/2026/11/ai-targeting-tests-iran-school-civilians/",
            "date": "2026-11-18",
            "data": "https://newmindmuseum.com/safety/2026/11/ai-targeting-tests-iran-school-civilians/data.json"
        },
        {
            "title": "AI models protected a peer from shutdown in tests",
            "url": "https://newmindmuseum.com/safety/2026/11/ai-models-peer-preservation-shutdown-study/",
            "date": "2026-11-14",
            "data": "https://newmindmuseum.com/safety/2026/11/ai-models-peer-preservation-shutdown-study/data.json"
        },
        {
            "title": "The UK AISI cybertesting incident: what happened, according to both sides",
            "url": "https://newmindmuseum.com/safety/2026/11/uk-aisi-cybertesting-incident-claude-mythos-5/",
            "date": "2026-11-06",
            "data": "https://newmindmuseum.com/safety/2026/11/uk-aisi-cybertesting-incident-claude-mythos-5/data.json"
        },
        {
            "title": "Sycophancy: why OpenAI rolled back a GPT-4o update",
            "url": "https://newmindmuseum.com/safety/2026/11/sycophancy-openai-gpt-4o-rollback-explained/",
            "date": "2026-11-06",
            "data": "https://newmindmuseum.com/safety/2026/11/sycophancy-openai-gpt-4o-rollback-explained/data.json"
        },
        {
            "title": "OpenAI and Apollo Research on &#8216;scheming&#8217;: what the study found, explained",
            "url": "https://newmindmuseum.com/safety/2026/11/openai-apollo-research-scheming-study-explained/",
            "date": "2026-11-06",
            "data": "https://newmindmuseum.com/safety/2026/11/openai-apollo-research-scheming-study-explained/data.json"
        },
        {
            "title": "Myth or fact: did Claude try to blackmail an engineer?",
            "url": "https://newmindmuseum.com/safety/2026/11/myth-claude-blackmail-engineer/",
            "date": "2026-11-06",
            "data": "https://newmindmuseum.com/safety/2026/11/myth-claude-blackmail-engineer/data.json"
        },
        {
            "title": "Who checks the AI labs? The rise of independent evaluators",
            "url": "https://newmindmuseum.com/safety/2026/11/independent-ai-evaluators-metr-anthropic-openai-aisi/",
            "date": "2026-11-06",
            "data": "https://newmindmuseum.com/safety/2026/11/independent-ai-evaluators-metr-anthropic-openai-aisi/data.json"
        },
        {
            "title": "Could AI systems hide their own mistakes? METR&#8217;s warning, explained",
            "url": "https://newmindmuseum.com/safety/2026/11/could-ai-systems-hide-their-mistakes-metr/",
            "date": "2026-11-06",
            "data": "https://newmindmuseum.com/safety/2026/11/could-ai-systems-hide-their-mistakes-metr/data.json"
        },
        {
            "title": "Agentic misalignment: Anthropic&#8217;s study of AI as an &#8216;insider threat,&#8217; explained",
            "url": "https://newmindmuseum.com/safety/2026/11/agentic-misalignment-anthropic-insider-threat-explained/",
            "date": "2026-11-06",
            "data": "https://newmindmuseum.com/safety/2026/11/agentic-misalignment-anthropic-insider-threat-explained/data.json"
        }
    ],
    "license": "https://creativecommons.org/licenses/by/4.0/"
}