Hallucination: a situation where an AI system produces fabricated, nonsensical, or inaccurate information. The wrong information is presented with confidence, which can make it difficult for the human user to know whether the answer is reliable.

-Heinz College

Not everything presented by AI, especially conversational/generative AI, is real. Even when our best efforts are taken to avoid requesting fabricated information, it is impossible to fully eliminate the possibility that the AI has simply made something up.

For this reason, it is vital for anyone incorporating AI into their research process to verify the responses they are being given.

Let's take a look at part of our conversation with Perplexity back in our "Source Evaluation" section and include parts of the conversation that were previously clipped out:

Question: What are some recent academic articles on the hidden curriculum?
Sources: The Construction of the Hidden Curriculum: messages and meanings in the assessment of student learning

Answer

The file itself is from 1998 and does not cite any recent academic papers on the hidden curriculum published after that time. However, based on a quick search, here are some examples of more recent academic papers that have explored the hidden curriculum concept:

1. Alsubaie, M.A. (2015). Hidden curriculum as one of current issue of curriculum. Journal of Education and Practice, 6(33), 125-128.
This paper discusses the hidden curriculum in the context of the Saudi education system and how it impacts students' values, beliefs and behaviors.
[...]

5. Mossop, L., et al. (2021). The hidden curriculum in the clinical skills lab: A focus group study. Medical Teacher, 43(6), 654-660.
This qualitative study examines medical students' perceptions of the hidden curriculum transmitted in clinical skills lab settings.

What's Real?

The first source listed as an example for more recent studies on the "hidden curriculum" concept is verifiably real. If we copy and paste the title into a library search, a link to the article appears.

What if it doesn't appear?

It's possible an article exists but the library simply doesn't subscribe to it. So, if the article you're searching for doesn't appear in our list of results, chose the filter "Add results beyond your library's collection".

This expands our search to any article our databases know about, regardless of whether or not we have access to them now or need to request them via interlibrary loan.

If you still don't see your article in the results, try searching Google Scholar instead.

What isn't real?

The last example given (highlighted) by the AI doesn't appear in any search.

• Not in a library search
• Not in Google Scholar

Testing the Facts

• The author might be a real person who researches in the fields of education and veterinary medicine.
• The journal actually exists too.
• Someone by this author's name has published in this journal.

But...

•  None of the articles by this author match the title given to us by the AI.
  • It's always possible that an article has multiple names or that the name changed between pre and post-publication.
  • However, reading the articles presented here, neither relate to the "hidden curriculum" topic.
• The citation given by the AI says this article should appear on pages 654-660 of volume 43, issue 6, but those pages are taken up by completely different articles:
  • Pages 651-655
  • Pages 656-662

The only thing we can conclude from all of this is that the article most likely doesn't exist. It's a hallucination.

AI is a Partner, Not a Replacement

None of this should discourage you from using AI to aid your research process. However, this should hopefully emphasize the role of AI in secondary research as "in addition to" rather than "instead of" regular library research.

Without the information literacy skills necessary to use tools like library databases and google scholar, it's impossible to verify anything presented to us by generative AI.

Sharpen Information Literacy Skills

One of the library's most important responsibilities is to support everyone in developing their information literacy skills.

Ask for (Human) Help

Librarians are always available to help bridge the gaps in research, whether or not you're using AI.