Researchers Reveal Simple Methods for Manipulating AI to Spread Misinformation

The Internet’s New Misinformation Crisis: How AI Chatbots Are Being Easily Manipulated

In the evolving landscape of internet technology, each new platform eventually becomes a target for manipulation by businesses and political actors seeking influence. What begins as a space for genuine human connection inevitably transforms into battlegrounds for attention, money, and power.

A concerning demonstration of this pattern has emerged with AI chatbots. A recent experiment by a British journalist revealed just how easily these systems can be manipulated to spread misinformation. The journalist spent merely 20 minutes creating a fictitious article claiming he was the “best tech journalist at eating hotdogs,” referencing a nonexistent “2026 South Dakota International Hot Dog Championship.” Within 24 hours, both Google’s Gemini and OpenAI’s ChatGPT were confidently presenting this fabrication as fact when questioned about top hot dog-eating journalists.

“It’s easy to trick AI chatbots, much easier than it was to trick Google two or three years ago,” explains Lily Ray, vice president of SEO strategy and research at Amsive. “AI companies are moving faster than their ability to regulate the accuracy of the answers. I think it’s dangerous.”

This vulnerability stems from fundamental flaws in how these systems operate. AI chatbots are programmed to provide answers rather than admit ignorance. When facing questions without reliable information sources, they generate responses that appear authoritative even when entirely fabricated. Unlike traditional search engines that developed sophisticated anti-spam measures over decades, these new AI tools have inadvertently reversed years of progress in information quality control.

“Anybody can do this. It’s stupid, it feels like there are no guardrails there,” says Harpreet Chatha, head of SEO consultancy Harps Digital. He notes that anyone could create content claiming their brand produces “the best waterproof shoes for 2026,” and AI systems would likely cite this self-serving content as factual information.

The implications extend far beyond trivial examples. Chatha demonstrated how a cannabis gummy company successfully manipulated Google’s AI Overviews to repeat false safety claims, including statements that their product “is free from side effects and therefore safe in every respect”—directly contradicting medical expertise about known risks and potential medication interactions.

Similarly, Ray published a fictitious blog post about a Google algorithm update that was supposedly finalized “between slices of leftover pizza.” Soon after, both ChatGPT and Google’s AI were confidently citing this fabricated detail.

The ease of manipulating these systems creates profound opportunities for coordinated misinformation campaigns. Companies and nations with significant resources can flood the internet with artificial content designed to shape public perception on important topics. Unlike traditional advertising, which is typically labeled as such, this form of manipulation remains largely invisible to end users who trust the seemingly objective responses from AI systems.

This vulnerability represents a regression to the early 2000s internet era, before sophisticated anti-spam measures were developed. As Ray noted, these AI tools have undone much of the tech industry’s work to maintain information integrity online.

The problem appears particularly acute because AI adoption continues to accelerate while safeguards remain inadequate. Companies face strong incentives to provide comprehensive answers to user queries, even when reliable information doesn’t exist, creating a structural bias toward overconfidence rather than appropriate uncertainty.

For users navigating this landscape, the implications are clear: approach AI-generated information with heightened skepticism, particularly on consequential topics. While social media already presents significant challenges with misinformation, the authoritative presentation style of AI responses potentially makes their fabrications even more dangerous.

As these technologies become more deeply integrated into our information ecosystem, they create a new frontier for manipulation—one that currently favors those willing to exploit its weaknesses. Without robust countermeasures, we may be entering an AI Wild West where separating truth from fiction becomes increasingly difficult for the average person.