Attachments
  • Mark as Completed
  • 10
  • More
Previous class
Section 1: Fact-checking

Section 2: Citing

Module 9 — AI for Research (Without Getting Misled)

Section 2: Citing

Purpose of This Section

This section explains how AI handles citations, why AI-generated references cannot be trusted by default, and how citation errors create risk in professional and organizational contexts.

AI can assist with research direction and synthesis, but it does not reliably verify or retrieve sources. When citations are incorrect, fabricated, or misused, credibility and trust are immediately compromised.

Citing is where research moves from exploration to accountability.

The Core Idea

AI can generate references, but it does not guarantee their existence or accuracy.

Generative AI systems produce text that resembles citations based on learned patterns. They do not consistently retrieve, validate, or confirm sources unless explicitly connected to verified databases and workflows.

A citation that looks real is not the same as a citation that is real.

How Citation Errors Occur

AI is trained on examples of how sources are commonly referenced. As a result, it may:

invent plausible-sounding studies or reports

combine real authors with incorrect titles

misattribute quotes or findings

format citations correctly while referencing nonexistent material

summarize claims without confirming original context

These errors are often subtle and difficult to detect at a glance.

Why Citation Errors Are Dangerous

Citations signal authority.

When a source is cited, readers assume the information has been verified and can be defended. Incorrect citations undermine credibility and can cause others to trust and repeat false information.

Once a faulty citation enters a document, deck, or report, it may be reused without rechecking, allowing errors to spread across teams and decisions.

The more official the format, the higher the risk.

Common Citation Failure Patterns

Citation-related issues often include:

studies that cannot be found when searched

laws or regulations cited inaccurately

outdated versions of policies or standards

statistics presented without verifiable origin

references that do not support the stated claim

These failures are especially costly in external communications, compliance-related work, and decision-making documents.

When Citations Matter Most

Extra care is required when citations are used to:

support recommendations or decisions

justify policy, legal, or financial positions

inform customers, partners, or stakeholders

document research findings or analysis

establish credibility or authority

When a citation is included, it implies responsibility for accuracy.

How to Use AI Safely When Citing

AI should be used to assist discovery, not to finalize references.

Responsible use includes:

using AI to suggest search terms or research directions

independently locating and opening original sources

confirming authorship, publication, and date

verifying that the cited source supports the stated claim

AI can accelerate research workflows.

Verification ensures that sources are defensible.

Common Failure Mode

A common mistake is assuming that a well-formatted or authoritative-sounding citation is reliable.

Another failure mode is copying AI-generated references directly into documents without verification, allowing fabricated or incorrect sources to be treated as legitimate.

Formatting does not equal validation.

The Conjugo Rule

Never cite what you have not verified.

AI can point you toward information.

You are responsible for standing behind the source.

Section Takeaway

AI-generated citations may be incomplete or incorrect

Plausible formatting does not ensure validity

Fabricated or outdated sources are common risks

Citations increase perceived authority

Verification is required before reuse or sharing

Responsibility for accuracy remains human

This concludes Section 2.