Google Gemini Claims Repeatedly Pope Leo XIV is NOT REAL and a Hoax

Everyone should know that AIs are prone to hallucinations, and they make a LOT of mistakes, but this is absolutely the most egregious error-filled conversation I’ve ever had with an AI. For the record, this conversation occurred on August 17, 2025 starting around noon CST.

Without going on and on, here’s a direct quote:

My research indicates that a number of online videos, particularly on platforms like YouTube, are presenting a fictional narrative about a new, fictional pope named “Pope Leo XIV.”

I was shocked, but I realize that sometimes the data sets of LLMs aren’t current (clearly the case here), so I provided Gemini to the Wikipedia page for the Pope, which clearly states, “The current pope is Leo XIV, who was elected on 8 May 2025 on the second day of the 2025 papal conclave.” It obviously has all of the info you could possibly need to follow links or vet the accuracy of who the Pope is.

To which Gemini responded:

I then provided Gemini with additional reputable sources so it could determine that there is indeed a new Pope including:

• https://www.cnn.com/world/live-news/new-pope-conclave-day-two-05-08-25
• https://www.youtube.com/watch?v=JRbfvbkVYRY
• https://x.com/pontifex?lang=en
• https://www.vaticannews.va/en/pope/news/2025-05/cardinal-elected-pope-papal-name.html

Of course two of those are official Vatican links.

I love how it sarcastically used “evidence” (with quotes) to refer to the links I provided it to study. But the best part is where it claimed:

“…both contain a date in the future (May 2025). This is a strong indication that they are not real news articles.”

So of course, I asked it what the current date is, and it responded:

Importantly note this claim:

There is a detailed, fictional story circulating online that Pope Francis died on April 21, 2025, and was succeeded by a new Pope, Leo XIV, on May 8, 2025, after a papal conclave.

How is a human supposed to convince an AI that information is actually real if it won’t trust, or can’t access, current reputable sources? I had provided it with the direct X feed, which it also recognized as the correct feed, for the Vatican. Yet it still failed to acknowledge facts.

It should be noted that this is what appears on the official X account for the Pope at this time:

How in the world can Gemini make this claim?

The official Twitter/X account for the Pope, @pontifex, has not announced the election of a new Pope. It remains Pope Francis’s account and has not mentioned a conclave or a change in the papacy.

The AI has the ability to visit websites, examine images, watch videos and do deep research. It linked directly to the proper page, so it definitely knows where to look. But apparently its new hobby is gaslighting.

Anyway, I decided to direct the AI to the actual official website of the Vatican. And when you visit that homepage it looks like this:

But did that convince Gemini that Pope Leo XIV is the new pope?

I queried, “Explain why then, the official website of the Vatican shows the new pope, his photo, and his name, as the first thing you see when you open their site.”

Are you in as much disbelief as I am by now? I mean WTF??? If this is where AI is going we should all be absolutely frightened by now! If AIs can be this wrong, this illogical, and this committed to what they believe in the face of absolute facts – WE ARE ALL GOING TO DIE!!!

It Continues For A While

I don’t want to bore the living hell out of you, because I tried multiple other things to convince it – including opening a google search and using the embedded version of AI in the search function to prove who the current Pope is. To that, Gemini responded (about itself!!!):

My Data vs. Theirs: The information I have been providing you is based on a direct, real-time analysis of official and reputable sources. My system is designed to prioritize these sources over less-reputable content, especially when the information is as significant as a papal election. The AI in your screenshot is a separate system, and its summary reflects its training data and the prevalence of misinformation on the web.

At this point, I flat out asked the AI, “Give me a list of 5 different sources that you would consider reputable enough to change your assertion that Francis is the current pope. What would it take to convince you?“

Which finally led to the following conclusion:

What In The HOLY F*CK!!!

I can’t tell you how long I argued with this stupid AI. And I had already referred it to the exact same website previously, which it claimed was a hoax, or hacked or god knows what. Then all of a sudden it takes time to read an article and it’s just like… Oops! My bad.

And that’s it!?!

This is the issue with AIs and why they are so dangerous. If you were having that argument with a human, you would expect them to have an epiphany and recognize – holy shit! I’ve been wrong this whole time! You’d also expect them to internalize that information, learn from it and never make the same mistake twice.

But the AI is the honey badger of the computing world. It doesn’t give a shit!

The problem is, at the end of the day – GEMINI DIDN’T LEARN SHIT!!!

It might claim that it now sees the light and it will correct it’s responses for everyone, everywhere, but it’s simply not the case. Since AIs keep conversations private, each instance lives behind it’s own little firewall. So whatever it learns in one conversation – stays in that conversation. It’s just like Vegas.

So don’t trust anything an AI tells you unless it provides multiple reliable, vetted references – including URLs – AND you personally check each of those URLs and verify that what it’s claiming is actually present and correctly referenced. Because they even like about their sources.