{"data":{"id":"38d703be-aea9-4589-9ef6-9dc44ca7b638","title":"Practical lessons from deploying AI securely at scale","summary":"Enterprise AI security challenges emerge not from model vulnerabilities but from how AI integrates into business workflows, where it accesses multiple systems and makes decisions autonomously. Traditional security controls focus on authentication (who the AI is) and authorization (what systems it can access), but fail to address what actions the AI should actually perform once it has access, creating gaps where authorized systems can act in ways that violate business intent. Organizations need runtime governance (monitoring and controlling AI behavior during execution) rather than just credential-based controls, because AI systems reason and generate unpredictable outputs that static security policies cannot adequately constrain.","solution":"N/A -- no mitigation discussed in source.","labels":["security","policy"],"sourceUrl":"https://www.csoonline.com/article/4205710/practical-lessons-from-deploying-ai-securely-at-scale.html","publishedAt":"2026-08-06T09:00:00.000Z","cveId":null,"cweIds":null,"cvssScore":null,"cvssSeverity":null,"severity":"info","attackType":[],"issueType":"news","affectedPackages":null,"affectedVendors":["Microsoft"],"affectedVendorsRaw":["Microsoft","Microsoft 365","SharePoint","ServiceNow","Salesforce","GitLab","Outlook"],"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":"2026-08-06T09:00:00.000Z","capecIds":null,"crossRefCount":0,"attackSophistication":"moderate","impactType":["integrity","confidentiality"],"aiComponentTargeted":"agent","llmSpecific":false,"classifierConfidence":0.85,"researchCategory":null,"atlasIds":null}}