Skip to Main Content
AUS Library Homepage
University Library

AI Literacy @ AUS

Evaluating Information from Generative AI

Generative AI tools produce text by drawing from patterns in their training data, but how they assemble information isn't always transparent. This uncertainty makes it challenging to assess the accuracy, reliability, and authority of what they generate.

Consider the following questions when evaluating AI-generated content:

  • Where is the information coming from?
    AI tools don’t cite sources consistently - or at all.

  • Can you identify the original authors or sources?
    AI may pull from unnamed works without attribution.

  • What is missing?
    Key voices, perspectives, or peer-reviewed sources might be excluded.

  • Are the citations real and accurate?
    AI tools can generate fabricated or incorrect references.

  • Is the content paraphrased or directly copied?
    AI may unknowingly reproduce copyrighted material.

  • Can the information be independently verified?
    Always cross-check claims with credible academic sources.

  • Has the content been peer-reviewed?
    Most AI-generated text is not based on peer-reviewed research.

Be skeptical, think critically, and recognize your own biases when evaluating AI-generated content. The responsibility for accuracy and academic integrity remains with you.

Echo Chambers and Algorithmic Biases

The internet is designed to personalize your experience. Platforms like Google, Instagram and TikTok use algorithms to prioritize content they think you'll engage with. While convenient, this customization creates echo chambers - environments where you're mostly exposed to ideas that reinforce your existing views. This can limit exposure to diverse perspectives and subtly reinforce personal and societal biases.

Don’t take content at face value. Use the SIFT method: 

These steps will help you seek out reputable sources, identify credible authors, and stay critical of misleading or biased content - especially in spaces where algorithms shape what you see.

Types of Bias

Generative AI reflects the biases in its training data, often reinforcing societal prejudices or amplifying certain viewpoints. Its opaque algorithms can unintentionally spread misinformation and skewed narratives. Recognizing these biases is essential to ensure more balanced and fair use of AI-generated content.

Overcoming Confirmation Bias