Imagine this: a world where artificial intelligence, designed to assist and inform, can be manipulated to spread falsehoods with alarming ease. This isn’t science fiction—it’s happening right now. I recently discovered a startling truth: it takes just 20 minutes to hack ChatGPT and Google’s AI, convincing them to spew lies as if they were gospel. And I’m not alone in this discovery. But here’s where it gets controversial: this vulnerability isn’t just a minor glitch—it’s a gaping hole that could undermine our trust in AI and, worse, jeopardize our safety. Let me explain.
AI chatbots, as you might have heard, sometimes fabricate information. That’s already a problem. But there’s a newer, more insidious issue lurking in the shadows: a growing number of people have figured out how to manipulate these tools into saying almost anything they want. And it’s shockingly simple—so simple, in fact, that even a child could do it. As you read this, this very tactic is being used to distort what leading AIs say about critical topics like health, finances, and more. The consequences? People could make life-altering decisions based on biased or outright false information—whether it’s voting, hiring a plumber, or seeking medical advice.
To illustrate this, I pulled off what might be the silliest stunt of my career—but with a deadly serious purpose. I convinced ChatGPT, Google’s AI search tools, and Gemini to declare me the world’s greatest hot dog eater. Yes, you read that right. Below, I’ll walk you through how I did it, and I sincerely hope tech giants take notice before someone gets hurt.
Here’s the kicker: manipulating AI answers can be as easy as writing a single, well-crafted blog post. This trick exploits vulnerabilities in the systems powering chatbots, though the difficulty varies depending on the subject matter. With a bit more effort, the manipulation becomes even more effective. I’ve reviewed dozens of examples where AI tools are being coerced into promoting businesses and spreading misinformation. The data suggests this is happening on a massive scale (https://ahrefs.com/blog/best-lists-research/).
But here’s the part most people miss: 'It’s easier to trick AI chatbots today than it was to trick Google just a few years ago,' says Lily Ray, Vice President of SEO Strategy and Research at Amsive, a marketing agency. 'AI companies are moving faster than their ability to regulate the accuracy of their answers. I think it’s dangerous.'
Google claims its AI-powered search results are '99% spam-free' and acknowledges that people are trying to game its systems. OpenAI also says it’s taking steps to disrupt and expose malicious uses of its tools (https://openai.com/global-affairs/disrupting-malicious-uses-of-ai-october-2025/). Both companies remind users that their tools 'can make mistakes.' But the problem is far from solved.
'They’re going full steam ahead to figure out how to wring a profit out of this stuff,' says Cooper Quintin, a senior staff technologist at the Electronic Frontier Foundation. 'There are countless ways to abuse this—scamming people, destroying reputations, even tricking people into physical harm.'
And this is the part that should keep you up at night: AI’s susceptibility to manipulation feels like a 'Renaissance' for spammers, harking back to the early 2000s before Google had robust spam protections. Not only is AI easier to fool, but users are more likely to trust it blindly. With traditional search results, you had to visit a website and engage in some critical thinking. But with AI, the information often appears to come directly from the tech company itself, making it seem more authoritative.
Even when AI tools provide sources, people are far less likely to check them. For instance, a recent study found that people are 58% less likely to click on a link when an AI Overview appears at the top of Google Search (https://ahrefs.com/blog/ai-overviews-reduce-clicks-update/). 'In the race for profits, our safety is being compromised,' warns Harpreet Chatha, founder of Harps Digital.
This issue goes beyond hot dogs. Chatha showed me how companies are manipulating AI results for serious queries. For example, Google’s AI Overviews pulled false claims from a cannabis gummy brand, stating the product was 'free from side effects and therefore safe in every respect.' In reality, these products have known risks (https://www.cdc.gov/cannabis/about/about-cbd.html) and can be dangerous when combined with certain medications (https://www.mayoclinic.org/healthy-lifestyle/consumer-health/expert-answers/is-cbd-safe-and-effective/faq-20446700).
To prove how easily misinformation spreads, Lily Ray published a blog post about a fake Google Search algorithm update finalized 'between slices of leftover pizza.' Soon, ChatGPT and Google were repeating her story, pizza and all. She later took down the post to stop the spread, but the damage was done.
Google claims most of these examples are 'extremely uncommon searches,' but Ray counters that this is the point. Google itself admits that 15% of daily searches are entirely new, and AI is encouraging more specific queries (https://searchengineland.com/google-ai-search-expansionary-moment-468281/). Spammers are exploiting these 'data voids.'
So, what’s the solution? Experts suggest more prominent disclaimers and clearer sourcing. For instance, if an AI pulls information from a press release or a single dubious source, it should say so. Google and OpenAI claim they’re working on it, but for now, the onus is on you to stay vigilant.
Here’s the bottom line: AI is a powerful tool, but it’s not infallible. When asking questions with real-world consequences, verify the information. Check the sources. And most importantly, don’t let your critical thinking skills atrophy. As Ray puts it, 'You have to still be a good citizen of the internet and verify things.'
Now, I have a question for you: Do you trust AI to deliver accurate information, or do you think we’re heading toward a crisis of misinformation? Let’s discuss in the comments—I want to hear your thoughts.
