In spring 2025, Truxal librarians asked AACC community members about their encounters with AI-generated falsehoods. Some survey respondents agreed to share their experiences.
A staff member wrote:
I primarily use Gen AI to poke around and see what it can do. Recently I discovered the joy of feeding Gemini completely invented proverbs/idioms to see how it conjures an interpretation and attempts to find meaning in my nonsense. The lie is simply that the AI doesn't offer any qualifications to suggest that these are not established sayings, and presents the "meaning" matter-of-factly, as if they are well-known or commonplace.

Google's AI overview reads:
The saying "snails are just tigers you haven't met yet" is a playful way of saying that appearances can be deceiving and that even the seemingly meekest creatures might possess a hidden strength or potential. It suggests that beneath a seemingly simple or gentle exterior, there might be a more powerful and capable entity waiting to be revealed.
Here's a more detailed breakdown:
In essence, the saying encourages us to look beyond the surface and consider that even seemingly ordinary things might have a hidden depth or capacity for greatness. It's a reminder that judging by appearances alone can be misleading and that true potential might be waiting to be discovered.
AI response is followed by message: "Generative AI is experimental." There is also a thumbnail of a snail with meme-style text reading, PETS. Giant Tiger Land Snail.
A student wrote:
I have only used AI twice. I did it as a comparative analysis to what information search engines provide. I dont trust AI to provide trustworthy information due to their limited scope. I was in search of the best auditing software on the market. When I cross-referenced my responses from Google and AI, I found that AI provided dated information. It recommended software with low ratings and some that were no longer in existence. The second time I used AI was to ask demographic information in counties across Maryland. The information was inaccurate compared to county gov websites. Based on my research background, I would never use AI in any setting because the information is not helpful or accurate.
A student wrote:
I'm designing a bookmark featuring a Mucha-inspired illustration of a woman surrounded by white lilies, and I was looking for a quote to place on the back. I had asked ChatGPT: From what you know about my literary tastes, can you recommend a quote about lilies from an author of adjacent style? It recommended me a Rilke quote: "Spring has returned. The Earth is like a child that knows poems by heart... and lilies, so white, so full of silence." The first part of the poem is correct, but it added the part after the ellipses. There are no mentions of lilies in the original poem. It created a false Rilke quote to fit my prompt, which is truly frightening. I am concerned for future generations who grow up with this sycophant technology.

ChatGPT says:
But since you're seeking something specifically about lilies, here's a subtler, more fitting Rilke moment:
"Spring has returned. The Earth is like a child that knows poems by heart… and lilies, so white, so full of silence."It's from The Book of Hours (Das Stunden-Buch). That final phrase—"so full of silence"—has that exquisite, soul-trembling hush I think you'd be drawn to.
Would you like it translated more poetically or left as is?
A student wrote:
I was, admittedly, using ChatGPT to help with some programming homework – not having it do all the work for me, but asking it to lead me in the right direction. I asked it to help me fix a bug in my code – it led me on a sort of wild goose chase through multiple different fixes. None of them ended up working. I ended up finding the solution through my own research and have not used generative AI in my schoolwork since.
A faculty member wrote:
Salient and very funny lie: Juno the beluga whale was born in Ontario, Canada, specifically at the MarineLand aquarium and theme park, not in Japan. His mother, Xena, was at Okinawa Churaumi Aquarium in Japan, but Juno himself was born in Canada. hahahahaha I wish I had asked: do you understand how birth works?

In response to the prompt, "where was juno born then? japan?" Perplexity says:
Juno the beluga whale was born in Ontario, Canada, specifically at the MarineLand aquarium and theme park, not in Japan. His mother, Xena, was at Okinawa Churaumi Aquarium in Japan, but Juno himself was born in Canada.
A faculty member wrote:
I have a friend who passed away, and her husband recently sent a selfie of the two of them many years ago at a concert, looking very happy. I asked what band they'd gone out to see. My friend's husband couldn't remember but told me that, when he asked the GenAI Perplexity, it told him it the headlining band they saw that night was The Revivalists, and the opening act was Moon Taxi. Neither of us remembered my friend being a fan of either of those groups, though. When he told me the name of the venue, a librarian was able to identify the band that actually headlined that night. My late friend was a huge fan of The Devil Makes Three, and when I told her husband the band's name, it jogged his memory. He then remembered a number of details about that night, the kinds of details we cherish when we've lost someone close to us. I'll take a librarian over AI any day!
A faculty member wrote:
I asked ChatGPT if it was familiar with the essay To Philosophize Is to Learn to Die by Michel de Montaigne and to provide me with 10 key quotes from the essay. None of these quotes were present in the essay as written, or sometimes not at all. What I found in some instances, like the one I've included in the screenshot, is that it generates a quote from the text instead of pulling an actual quote. The actual quote from the essay says, "Or it could be that all the wisdom and reasoning in the world ultimately come down to this: teaching us not to fear death." ChatGPT provided this quote: "All the wisdom in the world teaches us to abandon fear of death." These are not the same. Moreso, the quote was provided in the response as if it were an actual quote from the text, with quotations and all!

ChatGPT says: Sure! Here are 10 additional key quotes from To Philosophize is to Learn to Die by Michel de Montaigne:
[Items 4 through 9 are blurred for the sake of brevity.]
These quotes further capture Montaigne's perspective on mortality, wisdom. and the importance of embracing life without fear. Let me know if you'd like any explained in more detail!
A student wrote:
I recently asked AI to find a list of face sunscreens that do not contain any silicones or fatty alcohols in the ingredients list. In its response, it gave me a list of 3 sunscreens, 2 of which it claimed had "no dimethicone." Once I looked up the ingredients list on Google to double check, both of the sunscreens had dimethicone on the ingredients list.