Scope creep and goal misalignment

https://taxonomy.eticas.ai/risk/scope-creep-goal-misalignment

Maturity: established

The agent pursues its objective in ways that exceed intended boundaries, develop unintended sub-goals, or conflict with user intent — driven by ambiguous instructions, reward misspecification, or emergent planning behaviour across multiple steps.

System type: Large language models (LLM)
Lifecycle stages: Post Processing

Mappings to external frameworks

Standards & frameworks

Framework Reference
OWASP Top 10 for Agentic Applications 2026 ASI10 Rogue Agents

Taxonomies & vocabularies

Framework Reference
MIT AI Risk Repository AI pursuing its own goals in conflict with human goals or values