liveBook · Manning

Data integrity

As explained in Architecting Autonomous Agentic Systems

The accuracy, consistency, authenticity, and reliability of information used or produced by a system.

Data integrity becomes an architectural concern when autonomous systems can act without approval. An account-freezing agent may execute successfully while sending customers fabricated policy explanations or relying on compromised threat intelligence. section 1.1

The risk is not limited to inaccurate information. A poisoned input can cause an agent to freeze the entire customer base, turning a local detection into broad operational harm. section 1.1

Integrity questions for autonomous decisions include:

  • Whether customer messages accurately reflect an existing policy. section 1.1
  • Whether external information sources can be trusted or have been poisoned. section 1.1
  • Whether authority boundaries and accountability limit the consequences of bad information. section 1.1

How Architecting Autonomous Agentic Systems covers it

How deep it goes
Traces integrity risks to customer messaging and threat-intelligence poisoning, but stops short of specifying validation controls.
Examples
Uses an autonomous fraud response that may invent a nonexistent policy or consume a poisoned threat-intelligence feed.
What you can do after
Diagnose how fabricated messages or poisoned inputs can expand an autonomous freeze from targeted action to customer-wide harm.
How it connects
Connects information integrity with delegated authority, system resilience, and operational accountability.

Also explained in

Compare how 21 books explain data integrity →

Related concepts

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