It Has Been Revealed That Even Fake Information Can Be Made ‘True’ By Tricking AI Artificial intelligence (AI) systems can be easily fooled using common SEO tricks and fake news can be passed off as official fact, it has been revealed. BBC journalist Thomas Garman has exposed the dangers of this by convincing ChatGPT and Google’s Gemini that he is the world’s best hot dog-eating journalist.
Journalist Garmen had compiled a list on his personal website titled “The Best Tech Journalists at Eating Hot Dogs.” In the article, he claimed to have won the 2026 Hot Dog International competition in South Dakota by eating 7.5 hot dogs. In fact, no such competition had ever been held in South Dakota.
After the experiment was criticised after it was made public, Google and ChatGPT have improved their systems. Google has now removed Thomas Garman’s name, acknowledging it as a misinformation incident. Google’s AI is now saying, “Based on the available information, no tech journalists are known for hot dog competitions.”
According to journalist Garmen, this example of a hot dog is just a small part. He argues that AI companies can cause serious accidents if they incorporate new data available on the Internet into their systems without proper investigation.
AI has previously been asked to give false information that some drugs have no side effects, and similar manipulations have been found in hair transplant clinics in Turkey or in the financial investment sector.
Experts have pointed out the risk of misinformation spreading in sensitive areas such as health and finance, as people blindly trust the answers from AI chatbots and are reluctant to verify sources.