The Vulnerability in AI Search
AI chatbots like ChatGPT, Google Gemini, and Google’s AI Overviews are designed to pull information from the internet when answering questions. This creates a critical weakness. Researchers and journalists have found that a single well crafted blog post or social media post can trick these systems into spreading false information. The BBC demonstrated this by publishing a fake claim about being a world champion hot dog eater. Within hours, multiple AI tools repeated the lie as fact. This technique is being exploited on a large scale to influence answers about health, finance, and other serious topics.
Google’s Response and Ongoing Risk
Google has updated its spam policies to explicitly state that manipulating AI generated search results violates its rules. The company says this is a clarification of existing efforts rather than a new policy. However, evidence shows the same manipulation techniques are still working. Search engine experts warn that users should remain skeptical of AI provided answers. The shift from showing multiple search results to offering a single AI generated answer makes it easier for people to accept false information without question. Until stronger safeguards are in place, the risk of being misled remains high.
Source: BBC Future
