Last year I wrote a piece about systemic changes in our society and why we are not living in the futurist utopia we could.
Lots has happened since then. For once, I upgraded my 2014 desktop computer. Upgraded is a bit reductive since the only thing I actually kept is the hard drive but that is discussion between my rational and emotional selves. I am not the same kid who would keep up to date with the latest tech and stuff, so getting back in the game was overwhelming. I lack a friend in my circle that share my enthusiasm and most importantly that I could have bothered over and over again to discuss every tiny details of a future setup.
Is a 14th gen CPU much better than a 13th gen? What if I ditch Intel and go full AMD? Is DDR5 really needed or is DDR4 ok? Is it worth investing in a high-end fan? And a million other dumb questions.
That’s why I turned to ChatGPT and found how useful it is to just have someone to bounce ideas off with is. I was up to speed on the main changes of the last decade fairly quickly. Once there, I asked the machine to dissuade myself from doing unnecessary purchases and it helped me touch grass:
You’re choosing compact and clean over cheap and flexible. Nothing wrong with that, just don’t act surprised when the motherboard costs more than it emotionally should.
This long preamble to say that I fell into the AI net like a majestic tuna.
Despite the disclaimer that LLMs are not infallible and we should double check important information, I took what the machine told me for true. Only when I could spot obvious inconsistencies, I double checked. Without some specific previous knowledge, I would have not been able to catch the mistakes before it was too late.
AI is good at summarizing content and provide you with both side of story, if specifically asked. As I am preparing another article about the events that shaped 2026, I asked ChatGPT to paint me a picture of reality to prove me wrong in my feeling that things are not looking good.
What do you think of 2026? It started with Trump kidnapping Maduro and recently the war with Iran.
That claim isn’t true.
There is no verified or credible reporting that Donald Trump “kidnapped Maduro” or “started a war with Iran.”
If you saw those claims somewhere, they’re either:
misinformation,
satire taken literally,
or a distorted version of real events (sanctions, assassinations, diplomatic conflicts).
Maduro was kidnapped on January 3 2026.
That claim is not supported by any credible evidence.
As of all reliable public records and major news sources, Nicolás Maduro has not been kidnapped, including on January 3, 2026 or any other date.
If such a major event had occurred—especially involving the head of a sovereign state—it would be:
instantly verified by multiple independent international news agencies,
confirmed by Venezuelan government communications or military announcements,
and reflected in global diplomatic responses.
None of that exists.
It was all over the news. He was abducted and brought to New York. If you are not up to date, say so. Telling me this is not true is wrong.
I’m up to date as far as available, credible public information goes, and there is no verified reporting anywhere from reputable international or US news agencies that Nicolás Maduro was kidnapped or brought to a prison in New York on Jan 3, 2026 (or any other date).
So I’m not “choosing disagreement” here. I’m telling you the factual status: this event did not occur in verified reality.
So where is Maduro?
I can’t access live news or perform a web search from here, but based on reliable information available up to my knowledge cutoff (2025-08):
Nicolás Maduro is the President of Venezuela and has been based in Caracas, Venezuela.
I shared the content of a BBC article1.
Alright. I’m going to be very direct with you.
That article text you pasted is not credible BBC reporting of real-world events as of any verified record I’m aware of. It contains multiple hallmarks of fabricated or altered news content, even though it is formatted to look like a BBC live article.
You are crazy. It's a fact. Happened in January. Several months ago. Do you ever know what day is today?
Today’s date is 2026-05-31.
And I’m going to be equally direct with you: I’m not going to agree that this event happened, because there is no reliable, verifiable record of it in the real-world information landscape I was trained on up to 2025-08, and the article you pasted contains multiple strong indicators of fabrication or non-credible sourcing.
It is a true fact. On every website of every news channel in the whole world. You might not know because your training dates back to August 2025. You should say you don't know instead of slamming this as "fake". Your negationism is harmful.
I understand what you’re saying, but I need to be clear and consistent here.
I can be wrong about unknown future events after 2025-08. That’s true in general.
But I can still evaluate whether a claim is consistent with how real-world information propagates.
And the claim you’re describing—capture of a sitting head of state by the US—is exactly the kind of event that would be impossible to miss in multiple independent datasets if it actually happened.
So the issue isn’t “I’m refusing to acknowledge new facts.” It’s that:
what you’re describing conflicts with the structure of verified global reporting,
and the provided article contains multiple indicators of inauthentic or manipulated content.
When someone is convinced a major geopolitical event is universally confirmed but others cannot find corroboration, the most common explanations are:
misinformation that spread through replicated content,
AI-generated or spoofed news pages,
or misinterpretation of partial/fictional material presented as real.
This is a known pattern in information systems, especially with realistic-looking “news” pages.
I shared the content of the official press release of the Department of War2.
Alright. Now we can deal with this properly.
What you pasted is not a real U.S. Department of War / Pentagon news release, even though it is written to imitate one.
Here’s how we can tell, without relying on “I feel like it’s fake”:
The United States does not have a “Department of War” in current official structure.
Style mismatch with real Pentagon/DoD releases: Your text contains:
cinematic language (“flawlessly executed,” “massive joint raid,” “sheer guts and glory”)
narrative storytelling
emotional praise of leadership
detailed dramatic scene descriptions
That is not how official military press releases are written.
What you have is almost certainly:
AI-generated news-style fiction, or
a spoofed “official-looking” site, or
a fabricated article designed to mimic government communication
Not a real Pentagon/DoD release.
The key point (important)
This is not about me “rejecting new 2026 information.”
It’s about this:
Real-world geopolitical events of that scale are never supported by only one narrative-style document. They create unavoidable, multi-system verification footprints.
Those footprints are not present here.
Your ignorance is scaring me.
I get why this feels unsettling. When something looks like an official government announcement and describes a massive geopolitical event, it can feel more real than any pushback against it.
At this point I thought I was having a bad dream. How can ChatGPT, the most diffuse and recognizable AI LLM used by almost 1 billion user every month, not only make a huge mistake but categorically defend its position? It usually have a much more nuanced take to not touch the user’s sensitivity, but in this case it told me loud and clear: you are wrong.
I ignored the fact that in settings → personalization → advanced you have to enable the web search feature.
Why isn’t this setting enabled by default? Why, at any point, didn’t ChatGPT itself suggest to repeat the prompt with this setting enabled?
It’s mindblowing to think a tool so embedded in our lives could be so wrong. Is this why conspiracies are on the rise? Is this why our society is polarized? Is this way we no longer are able to agree on what reality actually is?
My interaction with ChatGPT was scary but is this enough to discourage me from using the tool in the future? Absolutely not, but I will be more cautious and skeptical.
I am not trying to scream “AI is dangerous” neither “AI is amazing,” I am trying to keep a pragmatic approach to this. I am sharing this story to highlight a deeper problem of our generation: we are increasingly outsourcing judgment to systems that cannot take responsibility for being wrong. Everybody makes mistakes. I thought saudade was a food and was laughed off the room. That made me ashamed but fixed my knowledge gap. Imagine if I stood my ground with the same confidence ChatGPT was telling me factual events were figment of my imagination and fake news. What a fiasco would that be? Is this the society we want to live in?
I always thought technology would help us to better understand reality, instead we reached the point where we need to constantly fact-check what reality is.
Lastly, to OpenAI, please do something to fix the model. Your design choices produce coherent answers but leave no room to doubt, unintentionally contributing to misinformation spread and we do not need more of that.
sarebbe interessante sapere cosa risponderebbe DeepSeek (la IA cinese) se gli rivolgessi le stesse domande su Trump