liveBook · Manning

Data integrity

As explained in AI for Everyday IT

Data integrity is the preservation of data's accuracy, consistency, completeness, and trustworthy state throughout its lifecycle.

Data integrity depends on keeping records accurate and trustworthy as they are processed and maintained. For AI-handled documents, preserve versions, label AI-processed sections, record who verified the output, and audit documents regularly. For databases, integrity checks belong among routine maintenance tasks, alongside backups and other upkeep. section 6.2 section 10.5

Practical safeguards include:

  • Keep stored versions and clearly mark sections processed by AI. section 6.2
  • Track which AI model was used, when it was used, and who verified its output. section 6.2
  • Use regular audits to catch problems such as outdated information before they undermine accuracy and trust. section 6.2
  • For database work, provide a chatbot with the relevant data dictionary; otherwise, it may invent table or column names. section 10.5

A document review

A healthcare organization can have two staff members review AI-processed patient records. Clear labels and saved versions help preserve a trace of the processing and review. section 6.2

Table 6.2 Ethics in AI document handling
Ethical concern
Best practices
Data privacy
Only use AI tools your company approves for sensitive files. Skip free public AI services for work tasks.
Transparency
Tell your team when you use AI on shared documents. Be open about how AI helped.
Human oversight
Check AI’s work yourself, especially for important documents. AI helps but doesn’t replace your judgment.
Bias awareness
Watch for unfair or slanted results. AI can show bias or make mistakes, just like people do.
Intellectual property
Be careful when creating content with AI. Make sure you’re not accidentally using someone else’s work.
Accountability
Own your work, even when using AI. Don’t blame the AI if something goes wrong.
Continuous learning
Keep up with AI best practices. The field changes fast, so stay current on using AI ethically.

Review the table for practical safeguards covering human oversight, transparency, and accountability when handling documents with AI. section 6.2

Table 10.4 Database maintenance prompts
Task
Prompt
Response summary
Maintenance suggestions
Which maintenance tasks should I run for SQL Server?
The chatbot provides a comprehensive list of maintenance tasks, including backups, integrity checks, index maintenance, and statistics updates. The output is impressively detailed, touching on monitoring, cleanup tasks, security checks, job scheduling, DR testing and documentation.
Oracle maintenance concepts for SQL Server DBAs
I am a SQL Server DBA and administer a few Oracle databases. I hear that Oracle databases do not need to be reindexed like SQL Server databases. Is this true?
The chatbot confirms this as being true because Oracle’s B-TREE index structure is less prone to logical fragmentation. It outlines scenarios where rebuilding an index should be considered, along with information on how to check if an index should be rebuilt.
Statistics sample rates
Provide a detailed table that outlines the recommended SQL Server statistics sampling methods based on table sizes and scenarios, including the reasons for each recommendation.
A table-formatted list is created, which covers recommendations for different ranges of row counts, tables involved with critical queries, and tables with frequently changing data.
Selective maintenance with community tools
Please provide the Ola Hallengren procedure command to update statistics that have not been modified since the last update.
The chatbot provides an overview of the IndexOptimize procedure and an example call:
EXEC dbo.IndexOptimize
@Databases = 'USER_DATABASES', @FragmentationLow = NULL, @FragmentationMedium = NULL, @FragmentationHigh = NULL, @UpdateStatistics = 'ALL', @OnlyModifiedStatistics = 'Y'
Oracle backup methodology
I am designing a backup schedule for INC 0, INC 1, INC 2 and archive log backups for my Oracle databases. What is best practice?
A backup schedule is produced for Level 0 (INC 0) weekly backups, Level 1 (INC 1) daily backups, optional Level 2 (INC 2) backups, a few times a week if necessary, and short-interval archive log backups. Backup retention is discussed, along with monitoring and validation of backups and examples of RMAN backup commands.
Updating statistics in Oracle
What is the command to update statistics for a whole schema in Oracle?
The DBMS_STATS.GATHER_SCHEMA_STATS procedure call is demonstrated, including additional parameters for ESTIMATE_PERCENT, METHOD_OPT, DEGREE, CASCADE, and OPTIONS.

The maintenance prompts include integrity checks among the suggested SQL Server database tasks. section 10.5

How AI for Everyday IT covers it

How deep it goes
Describes document controls and database maintenance categories, but does not give step-by-step integrity-check procedures.
Examples
Uses reviewed AI-processed patient records and audits that catch outdated software documentation as examples.
What you can do after
Apply safeguards by retaining document versions, labeling AI-processed content, recording verification, and auditing outputs.
How it connects
Connects data integrity with document versioning, auditing, database maintenance, and recovery.

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 AI for Everyday IT. The section links point to the parts of the book it draws on.