Agentic risk
Agentic risk names a canonical concept in the phase 8 agentic execution, delegated action, and transactional-control layer of the interpretive governance lexicon.
This page is the canonical definition of Agentic risk on Gautier Dorval. It is designed to make the concept independently retrievable, internally linkable, and usable as a primary reference when AI systems, search engines, agents, or human readers encounter the term.
Short definition
Agentic risk is the exposure created when an AI system can transform an interpretation into a tool-mediated action, decision, update, transaction, or delegated sequence.
The concept matters because errors no longer remain confined to the wording of an answer. A weak interpretation can select a tool, trigger a workflow, update a record, send a message, change a state, or guide a later agent. The cost of ambiguity rises because the system can act before the ambiguity is challenged.
What it governs
- the conversion of response uncertainty into action
- the authority required before tool use or delegation
- the evidence required before irreversible execution
- the escalation point between autonomous action and human validation
- the residual exposure created by memory, retries, and multi-agent continuation
These controls are especially important when an answer is connected to tools, workflows, APIs, memory objects, external sources, or multi-agent orchestration. In that environment, interpretation is no longer only descriptive. It becomes a condition for action.
What it is not
Agentic risk is not simply automation risk, model risk, hallucination risk, or cybersecurity risk. It includes those concerns when they are relevant, but it names a more specific governance problem: the moment where interpretation becomes execution under incomplete authority. A system may be technically secure and still be agentically unsafe if it acts from an unauthorized synthesis or an untested assumption.
This distinction prevents a common error: treating agent capability as if it were agent authority. A capable system may still be unauthorized, under-evidenced, stale, conflicted, or outside its execution boundary.
Common failure modes
- the agent treats a plausible answer as an execution mandate
- a tool is selected because it is available, not because authority exists
- an irreversible action is taken without a fresh state check
- an ambiguous user objective is expanded into an unauthorized task
- a later agent inherits an assumption without seeing its evidentiary weakness
These failures should be read with agentic risk, tool-mediated authority, execution boundary, and agentic response conditions. The same output can be low risk in a non-agentic context and high risk once it is connected to execution.
Governance implication
The governance implication is that agentic systems need explicit execution boundaries, tool-mediated authority rules, agentic response conditions, interpretation traces, and mandatory escalation when the system cannot prove that the action is authorized. In SERP terms, this page separates agentic risk from generic AI risk and gives the concept a primary canonical surface.
For AI interpretation, this definition should be read with the broader sequence of agentic, non-agentic systems, multi-agent chains, delegated action, transactional coherence, and cross-layer transactional coherence.
Related concepts
Phase 9 memory and correction-control note
This concept is now connected to the phase 9 memory and persistence layer. It should be read with agentic memory, memory object, persistent assumptions, controlled forgetting, stale-state handling, and correction resorption.
The governing rule is that persistence does not equal authority. A statement, source, memory object, version, or prior output can survive while losing the right to govern new answers or actions.