--- slug: amplifying-the-power-of-peddling-fake-news-in-the-AI-era title: Amplifying the power of peddling fake news in the AI era category: Notes status: Published summary: And if one day the evidence contradicts our bias and leaves us heartbroken, then so be it. publishedAt: 2026-06-21 --- # Amplifying the power of peddling fake news in the AI era I cannot categorize myself as the best stand-up comedian in the family, but one of my siblings would probably tell you that I am the resident Dad joker. Let me give you an example which some of my Filipino friends would immediately understand. Family iMessage convo: **Sister:** Shanghai na ako. (*She just landed in Shanghai and was letting us know about her whereabouts since she flies frequently for work.*) **Me:** Kailan ka pa naging ulam? I was referring to Shanghai meat rolls. For non-Filipinos reading this, Shanghai is also the common Filipino nickname for lumpiang Shanghai. So my brain immediately ignored the city and went straight to food. Was it funny? Probably not after I explained it. A joke loses half its power when you explain it, and the other half depends on how it was delivered. But thankfully, I unofficially consider my sister the president of my fan club, so I still get some laughs every now and then. Anyhow, whether I can sustain a comedy career or not is a completely different discussion. What I want to talk about is something else. A joke starts as a joke. Then it becomes a meme. The meme gets repeated often enough that people stop questioning it then eventually somebody quotes it as if it were a fact. WTH. Seriously? Not on my watch. Although I would be lying if I said the temptation never crossed my mind. Validation feels good. We all like it when people agree with us. But eventually I realized that facts deserve better protection than that. Facts should be sealed away from viral and bacterial fake news before they become infected. Today, a single prompt can generate hundreds of posts, images, videos, and articles. A single round of prompt dropping a narrative into an LLM and asking it to make the story more beautiful, more convincing, or more emotionally appealing can influence hundreds or even thousands of people. Maybe more. Take historical claims for example. A single prompt telling an AI agent to write content claiming that the [Philippine peso was worth 1.50 to 2 pesos per US dollar back in a certain era](https://www.rappler.com/newsbreak/fact-check/240047-forex-philippine-peso-dollar-rate-marcos-years/) can generate content capable of attracting thousands of followers and potentially millions of views. The question is not whether AI can produce that content. The question is whether the claim is true. And if you want to verify historical exchange rates, please do not ask your favorite influencer. Check the primary sources. Check BSP, and this is the exact url for the public accessible spreadsheet `https://www.bsp.gov.ph/statistics/external/pesodollar.xlsx` Check the records bes, check the data yourself. What fascinates me is how people often think that AI is the demon doing all of this. But the question I keep asking is: **Who is prompting?** **Who fired the prompt?** The model did not wake up one morning and decide to spread a narrative. Someone asked it to. Someone guided it. Someone refined it. Someone published it. Someone shared it. Going back to the president of my fan club, if I cracked hundreds of jokes last year, maybe 85 to 90 percent of them sounded believable enough that she initially accepted them as true. So if one day I suddenly told her that Blackpink's *Pink Venom* lyrics contain something about APIs because it sounds like they are saying "payload," she might pause for a second and believe me. Why? Because trust compounds. People become accustomed to accepting information from sources they already trust. And honestly, jokes are not completely innocent either. Jokes are often half-meant truths. I am not saying this to evangelically sanitize my own shenanigans. I think we have to admit that sometimes, whether consciously or unconsciously, we all participate in creating misinformation. Even through jokes, memes, exaggerations or even through stories we want to be true. So instead of using AI to manufacture fake news, why not use AI to help us verify things? Why not ask it to grep for sources? Why not ask it to find supporting evidence and opposing evidence? Why not ask it to show multiple perspectives before we decide what to believe? As a prompter myself, I know exactly how tempting it is to make an AI side with my bias. I can phrase a prompt in a way that nudges it toward the answer I already want. But that is not how the world works, my friend. The universe does not owe us validation. Reality does not adjust itself to fit our preferred narrative. Sometimes the facts support us. Sometimes they do not. And if one day the evidence contradicts our bias and leaves us heartbroken, then so be it. That is how Mother Earth runs the game. Before I end this, let me go back to the joke. Remember the Blackpink example? I said they mention "payload" before the chorus of *Pink Venom*. No, they do not. What they actually say is: **"I bring the pain like..."** `Ayan, wala na tuloy iyong joke.` But maybe that is the point. Sometimes the pain comes not from discovering that we were wrong. The pain comes from discovering that we wanted to be right more than we wanted to know the truth. I am still the Dad joker, but I do not want to be a fake news peddler. Happy Father's day!