liveBook · Manning

Intended purpose

As explained in AI Governance

Intended use is the purpose for which a system is meant to be used, as specified by its provider or deployer.

Intended purpose is the use a system is meant for, including its specific context and conditions. It helps determine risk classification: systems using similar technology may fall into different categories because they serve different purposes. section 7.2

Reassess the intended purpose when actual use diverges from the approved scope. Signals can include:

  • New input or output types, tool calls, user groups, or downstream decisions. section 7.2
  • A new use that makes a system high-risk, or a substantial modification that changes its purpose or compliance in an unforeseen way. section 7.3.2

Purpose changes the classification

ClaimAssist began as a customer FAQ bot with predefined answers and no claims data access. Adding payment-plan eligibility assessments based on claims and payment data changed its purpose and risk classification. section 7.2

A deployer that changes a system’s intended purpose may become its provider if the new use is high-risk. Deployer evidence should show that the system was configured for its intended use and operated as designed. section 7.3.2 section 7.4.3

Figure 7.3 Classification and role decision tree. The tree produces a risk classification output

Use the classification decision tree to situate intended purpose within the system’s risk-classification workflow. section 7.2

How AI Governance covers it

How deep it goes
Defines intended purpose and traces classification changes through use shifts and monitoring signals, but leaves case-by-case judgments to documented reasoning.
Examples
Contrasts a customer FAQ with payment-plan eligibility, vendor-contract review with employee scoring, and assisted credit scoring with automatic approvals.
What you can do after
Evaluate a system’s specified and actual uses to identify classification changes and possible provider responsibility.
How it connects
Connects AI system classification, risk categories, provider and deployer roles, substantial modification, and monitoring.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of AI Governance. The section links point to the parts of the book it draws on.