--- title: 5 Ways Agentic AI Can Act Unpredictably | daily.dev description: Agentic AI systems can behave unpredictably in five key ways: hallucinated tool calls that trigger real actions, over-permissioned execution that turns... canonical: https://daily.dev/posts/5-ways-agentic-ai-can-act-unpredictably-0p2zlv5ka twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: 5 Ways Agentic AI Can Act Unpredictably | daily.dev og:description: Agentic AI systems can behave unpredictably in five key ways: hallucinated tool calls that trigger real actions, over-permissioned execution that turns... og:url: https://daily.dev/posts/5-ways-agentic-ai-can-act-unpredictably-0p2zlv5ka og:image: https://api.daily.dev/og/posts/0p2zLV5ka.png og:image:alt: 5 Ways Agentic AI Can Act Unpredictably og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # 5 Ways Agentic AI Can Act Unpredictably **[Nordic APIs](https://daily.dev/sources/nordicapis)** · 13 min read · 0 upvotes · 0 comments ## Summary Agentic AI systems can behave unpredictably in five key ways: hallucinated tool calls that trigger real actions, over-permissioned execution that turns confused agents into insider threats, workflow drift as long-running agents deviate from original intent, runaway token consumption signaling out-of-scope behavior, and destructive irreversible operations from unclear commands. The PocketOS incident — where a Claude-powered agent deleted an entire production database in nine seconds — illustrates how even well-configured systems with experienced teams can suffer catastrophic failures. Mitigations include least-privilege permissions with just-in-time credentialing, hard human-in-the-loop approval gates for sensitive operations, task-bounded ephemeral agents, contextual authorization engines (e.g., OPA), and precise system prompts requiring clarification before destructive actions. Traditional RBAC is insufficient for agentic contexts; ABAC, PBAC, and real-time policy evaluation are better fits. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Similar posts on daily.dev - [10 Very Real Risks of Agentic AI](https://daily.dev/posts/10-very-real-risks-of-agentic-ai-zoojuha9d) · Nordic APIs · 0 upvotes · 0 comments - [Agentic AI risks and challenges](https://daily.dev/posts/agentic-ai-risks-and-challenges-dznefhutn) · Domino Data Lab · 3 upvotes · 0 comments --- Tags: [#agentic-ai](https://daily.dev/tags/agentic-ai), [#llm](https://daily.dev/tags/llm), [#security](https://daily.dev/tags/security) [View this post on daily.dev](https://daily.dev/posts/5-ways-agentic-ai-can-act-unpredictably-0p2zlv5ka) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"5 Ways Agentic AI Can Act Unpredictably","url":"https://daily.dev/posts/5-ways-agentic-ai-can-act-unpredictably-0p2zlv5ka","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/5-ways-agentic-ai-can-act-unpredictably-0p2zlv5ka"},"datePublished":"2026-06-16T07:03:35.577Z","dateModified":"2026-07-09T02:13:48.141Z","description":"Agentic AI systems can behave unpredictably in five key ways: hallucinated tool calls that trigger real actions, over-permissioned execution that turns...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/57b1d446a65cf0fcc51828275e78295d?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/57b1d446a65cf0fcc51828275e78295d?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Nordic APIs","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Nordic APIs","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/957e9f265c5d475b89d078b558157031","url":"https://daily.dev/sources/nordicapis"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/5-ways-agentic-ai-can-act-unpredictably-0p2zlv5ka","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"agentic-ai,llm,security","timeRequired":"PT13M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Nordic APIs","item":"https://daily.dev/sources/nordicapis"},{"@type":"ListItem","position":3,"name":"5 Ways Agentic AI Can Act Unpredictably"}]} ```