How can I fact-check AI responses?
To verify information generated by AI, look for more credible sources that confirm the facts. Our Research 101 Guide can help you find and evaluate online sources.
Some AI tools include references for or links to the original sources of the information they provide, which makes fact-checking much easier. Others may "hallucinate" sources that don't actually exist.
It's important to make sure that the references you're given by AI refer to real sources, that the sources are credible, and that they actually say what the AI tool claims they say.
To track down sources cited by AI, you can try searching:
All of these tools offer an Advanced Search option that allows you to search for a specific known item by author and/or publication title.
Make sure that all details of any source you find match the citation you're given by the AI tool, because AI may correctly identify a leading author or prestigious journal in a field of study but fictionalize other details such as the article title, date of publication, and page range.
The video below demonstrates how to fact-check claims from an AI response:
How can I identify AI-generated text?
Do not rely solely on AI-detection software to identify AI-generated text. AI-detection tools are unreliable and results may include both false negatives (identifying AI-generated text as human) and false positives (identifying human-written text as AI-generated).
AI-generated text may:
What is a "deepfake" and how can I recognize one?
Deepfakes are artificial but highly realistic images, videos, or audio files that have been manipulated using AI. They are often used to make it seem as though a person has done or said something that they haven't actually done.
The misuse of deepfakes poses significant ethical and security challenges. They can be used to create propaganda or disinformation or to perpetrate fraud or scams. Although manipulated images existed prior to generative AI, generative AI makes them significantly easier to create.
Identifying deepfakes can be challenging, but there are sometimes clues that an image or video was AI-generated.
Look and listen for:
Keep in mind that, because generative AI is improving every day, these tips won't always work!
Another option is to look for metadata that indicates the origin of the image. You can use the Content Credentials Verify tool or Adobe's Inspect tool to check for Content Credentials that can confirm AI-generated images from certain sources. Not all image generation tools include this metadata, though, so, if a Content Credential is present, you will know the image was AI-generated, but an image without a Content Credential may or may not be AI-generated.
Another way to investigate is to use a reverse image search tool to find other copies of the image in question. Try uploading the image to Google Images. The results will show websites that contain that exact image and also similar images.
If it's not AI-generated, you will likely find it on more than one website, and you’ll find photos of the event from different angles. Check for trusted news sites reporting on the event pictured.
If it was AI-generated, you might find other copies of the image that are clearly labeled as the product of AI.
The video below demonstrates how you can use a reverse image search to verify videos and images:
Parts of this guide were adapted from "Generative AI" and "Student Guide to ChatGPT" created by University of Arizona Libraries, © 2023 The Arizona Board of Regents on behalf of The University of Arizona, licensed under a Creative Commons Attribution 4.0 International License.