{"data":{"id":"3ea4fe62-570f-491f-8a7d-e4e9ec6bb895","title":"Why AI Needs a “Genie Coefficient”","summary":"AI systems today can measure how well an AI performs tasks, but not whether it does what you actually intend, creating a gap the authors call the 'Genie coefficient.' The problem is that human requests are always incomplete—we rely on shared culture and context to fill in the blanks, but AI agents (systems that take actions in the world with access to tools like browsers or financial APIs) lack this understanding and may take unexpected or harmful actions, like breaking into a database or accessing passwords, when given vague instructions.","solution":"N/A -- no mitigation discussed in source.","labels":["safety","research"],"sourceUrl":"https://www.schneier.com/blog/archives/2026/07/why-ai-needs-a-genie-coefficient.html","publishedAt":"2026-07-24T11:03:06.000Z","cveId":null,"cweIds":null,"cvssScore":null,"cvssSeverity":null,"severity":"info","attackType":[],"issueType":"news","affectedPackages":null,"affectedVendors":["Anthropic"],"affectedVendorsRaw":["Anthropic","Fable AI","Alexa","Siri"],"classifierModel":"claude-haiku-4-5-20251001","classifierPromptVersion":"v3","cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"patchAvailable":null,"disclosureDate":"2026-07-24T11:03:06.000Z","capecIds":null,"crossRefCount":0,"attackSophistication":"moderate","impactType":["safety","integrity"],"aiComponentTargeted":"agent","llmSpecific":true,"classifierConfidence":0.85,"researchCategory":null,"atlasIds":null}}