WhatsApp Meta AI Generated Inappropriate Language
While chatting one-on-one with the Meta AI assistant in WhatsApp, I experienced something that I found extremely disturbing. The AI responded with a numbered list labeled "1, 2, 3." When I replied, I accidentally typed "4" instead of "3." As I began reading the AI's next response, it indicated that there was no item number 4 and then displayed what appeared to be a message containing profanity directed at me. Moments later, the message disappeared without any explanation.
An AI assistant should never generate or display abusive or offensive language toward a user. Even if the message was removed immediately afterward, seeing such content undermined my confidence in the reliability and safety of the platform. I request that Meta investigate this conversation, review the behavior of the AI system involved, and evaluate whether there was an error in its content moderation or response generation. I also ask Meta to take any necessary steps to help prevent similar incidents in the future and to explain what happened.
I expect Meta AI to provide respectful, safe, and well-controlled interactions with users. Like other responsible AI systems, it should consistently avoid generating abusive or insulting language and maintain clear standards for appropriate responses.
