The Complete Guide to AI Fact Checking For Students
byTrey Edgington
5 min to read
Generative Artificial Intelligence, or gen AI, can be a great information-gathering tool for research projects and other assignments, but educators, students, learning coaches, and parents still have a lot of questions. Is using AI cheating? Does AI provide quality research materials? How can students know that the information AI provides is accurate?
Regardless, if students are using AI for schoolwork or even generally, a large part of any online research is fact-checking AI-generated information for accuracy. Use this guide to learn more about fact-checking AI and how to use AI search tools smartly.
How Are Students Using AI for School and Research?
As ChatGPT and other large language models (LLM) gain popularity, more students are using them as search engines or as study aids. Google and Bing have added AI-generated answers to search results, which automatically summarize web pages based on keywords.
However, as good as this may sound, these AI summaries have been famously, and sometimes egregiously, wrong or presented without the proper context. A recent example includes the suggestion that people eat small rocks as part of their daily diet and adding glue to pizza sauce to prevent cheese from sliding off the pizza.
When using AI search results, students must remember to fact-check AI summaries for authority, context, and accuracy. Many AI search summaries come with sources. By checking those, you can more easily identify where the information came from and the context around the statement. This can help students more easily verify reputable sources.
Why You Should Start Fact-Checking AI Search Results
The first thing to do when looking at AI-generated search results, is ask yourself if something looks off about the information provided. It doesn’t have to be as egregious as eating rocks or putting glue in pizza. It could be something as simple as misinformation about George Washington driving a car—though that’s pretty bad! If it looks off, even in the slightest, it’s a good idea to double-check against a reliable source by tracking down where the information came from and deciding if the original source is reputable.
AI Search Results Can be Wrong
Many experienced editors and proofreaders will start by checking quotes and numbers, including dates, statistics, measurements, etc. They are the easiest places to start, and possibly the easiest to get wrong. Not only are numbers and quotes easy to get wrong, but wrong quotes and statistics can really ruin a school project, essay, or even a piece of professional work.
Imagine you’re doing research about the 1942 film Casablanca for a paper you’re writing for film class. The first response from Google may be an AI-generated summary of the film.
You ask for the most famous quote in Casablanca and the first answer is: “Play it again, Sam,’ which is likely the most famous quote from Casablanca. Humphrey Bogart’s character says the phrase to his piano player, referring to the song ‘As Time Goes By.’”
Though this answer may sound correct, it isn’t. No one in Casablanca actually says, “Play it again, Sam.” Instead, Bogart’s character says, “You played it for her, you can play it for me!”
While misremembering a quote from an old movie can be frustrating, apply the same instance to summaries misquoting a scientific research study or repeating a common misconception or confusing a fictional portrayal of a historical figure for what happened in real life. Only by checking AI search results against multiple reliable sources, can you find out the correct information.
Why Does AI Get Information Wrong?
The training data on many of these AI models comes from sources like user-generated websites that aren’t always reliable, peer-reviewed, or fact-checked. In the case of Google’s suggestion to eat rocks, the model picked up on an article from The Onion, a parody website, which often appears high up in search results and usually gets a lot of audience engagement, all things Google’s AI tool Gemini uses when creating its summary.
How to Fact-Check Your AI Content
Checking sources provided by AI will follow the same process as checking any other source, whether they be from interviews and websites to books, magazines, and scholarly journals. Students should ask themselves the following questions:
Does the author have the necessary credentials (i.e., education, professional experience, and reputation) to speak with authority on the topic?
Does the author cite their sources, and are these informed/peer-reviewed/biased?
Is the publication considered reliable?
Maybe most importantly, student researchers and Learning Coaches should trust their intuition; if something seems off, look a little deeper into information about the source.
Is AI Here to Stay?
We can look at AI as an educational helper, a way to cheat, or somewhere in between. When educators, students, and parents use it, we must all be mindful of the accuracy of the information we receive in return. The good news is that fact-checking AI is not much different from fact-checking other sources and information, so we don’t need to be intimidated; you already know how to do it! The important thing to remember when using AI search results is to check your sources and not to take anything at face value.
No matter what decisions you make, AI might just be here to stay, so it can be an important new computer skill to add to your educational repertoire.
Ready to Learn More
About Connections Academy?
Explore the benefits of attending Connections Academy, a tuition-free, accredited online public school that’s passionate about helping your child thrive.
