

Google's AI search results recently suggested adding glue to pizza sauce to prevent cheese from sliding off, based on a misunderstood joke from Reddit. Despite warnings, a curious individual tested this bizarre advice using non-toxic glue, resulting in a mediocre pizza that did keep the cheese in place. This incident highlights concerns about the reliability of AI-powered search results and the potential impact on public trust in such technologies.

Google's AI search engine mistakenly interpreted a Reddit joke about adding glue to pizza sauce as a legitimate cooking tip due to its inability to discern the humorous intent behind the comment5. The original Reddit post jokingly suggested using glue in pizza sauce to prevent the cheese from sliding off, a suggestion clearly not meant to be taken seriously. However, when queried about how to keep cheese from sliding off pizza, Google's AI, having ingested this comment, provided it as a viable solution. This incident highlights the challenges AI faces in understanding context and distinguishing between serious advice and humor, particularly when processing information from diverse and informal sources like Reddit5.

Google's AI has provided several absurd answers, as noted in the article. Some of these include:
These examples highlight the current limitations and errors in Google's AI-powered search results, leading to user mistrust in AI-generated answers.