Checking the credibility of sources used in your work ensures that information is accurate, reliable, and free from misinformation. Verifying sources also strengthens your arguments by backing them with trustworthy and peer-reviewed evidence. It also helps maintain academic integrity, supports well-informed decision-making, and prevents the spread of false or biased narratives.

This is especially important for Generative AI output, as GenAI can sometimes produce outdated, incorrect, or fabricated information that needs verification with credible academic sources.

To check the content of GenAI generated content follow the steps below:

1. Identify key claims and facts

Break down the AI-generated content into key statements and identify specific facts, statistics, or references that require verification.

2. Cross-check with reliable academic sources

You must check the ideas in the GenAI text are true. To do this identify the facts you want to check, and do a search in academic databases to see if you can verify the information. You can use the JCU OneSearch library database, or Google Scholar, or one of the other discipline specific databases available here. If you can not verify the information you must consider the GenAI output unreliable.

3. Look for consensus in scholarly literature

If multiple peer-reviewed studies support the GenAI's claim, it is likely credible, but be cautious of claims that lack supporting evidence or rely on a single study.  If you only find one or two academic papers that support the claim, you will need to check the credibility of the academic source.

4. Evaluate the credibility of academic sources

For all sources that you use, you must check factors including authority (the author's expertise on the topic), publication (whether it is from a peer-reviewed journal or reputable institution), date (the relevance and recency of the information), bias (whether the source has a particular agenda), evidence (if the source provides citations or references to support its claims), and methodology (the source provides a clear explanation of the data collection methods, the analytical techniques used, and any limitations of the study). If the source does not meet these criteria, it may not be a credible academic source.

5. Check citations provided by GenAI Output

Sometimes GenAI will provide citations for the information it presents. However, GenAI is known to create inauthentic references. If GenAI provided a source you must still authenticate that source. Verify it by searching for the exact paper in an academic database. If you cannot find the paper, it is likely inauthentic, and you cannot rely on the information generated. If you do find the source,  make sure you evaluate the credibility according to point 4 above.

There is more information on evaluating sources on the JCU library page here.

SIFT Method

The SIFT method is a way of reading a source critically. The SIFT method helps you get better at distinguishing truth from fiction, focusing on relevant details in a text, and understanding what a well-reasoned discussion looks like. It will help you to develop critical thinking skills, practice media literacy, and engage with credible sources. We can use this method when we are reading GenAI produced content as well as human generated content.

_{Image source:  https://hapgood.us/wp-content/uploads/2021/02/sift-infographic.png}

You can use the following steps to analyse GenAI output for credibility.