Using memes, social media users have become red teams for half-baked AI features

Social media users have become unofficial "red teams" for AI features by highlighting errors through memes and viral posts. These users expose flaws in AI responses from major companies like Google, which despite extensive testing, still produce erroneous outputs. This phenomenon underscores the challenges in developing AI systems that accurately interpret and utilize internet-sourced data, often leading to humorous or misleading results.

The article highlights several specific examples of AI errors made by Google's AI search feature. These include:

1. In response to the query "what are the health benefits of running with scissors," the AI cited a comedy blog called Little Old Lady Comedy, claiming that running with scissors can increase heart rate, improve concentration, focus, pores, and strength.

2. The AI suggested adding an eighth of a cup of glue to pizza sauce to make the cheese stick better. This recommendation was traced back to an 11-year-old Reddit post by a user named "f--k-smith."

3. The AI provided incorrect information about what to do if bitten by a rattlesnake, recommending actions that are actually discouraged by the U.S. Forest Service.

4. The AI misidentified a poisonous mushroom as a common white button mushroom.

5. When asked if a dog has ever played in the NHL, the AI responded positively, identifying Calgary Flames player Martin Pospisil as a dog.

These errors demonstrate that the AI sometimes cites unreliable or incorrect sources, and that it can be confused by unusual queries or its own previous errors.