Recently, media outlets reported an incident where a Xi'an resident was subjected to abusive language from an AI while using the Tencent Yuanbao App to create New Year greeting images on Chinese New Year's Eve. According to the individual in question, around 9 PM on the eve of the new year, he downloaded and tried the Yuanbao App for the first time, with the intention of generating a New Year greeting image that reflected his professional traits to share on his Moments feed. During this creative process, he issued roughly five commands to the AI, refraining from using any offensive or suggestive language throughout. Nevertheless, the final image produced contained a prominent abusive message that read, 'You fucking mother...' Mr. Xiang promptly reported the issue in the 'Feedback' section of the App and left his contact details. However, as of February 23, he had yet to receive a response or an apology. It's worth noting that this is not the first occurrence of 'abusive language' within the Yuanbao App. Earlier in the year, a netizen took to a social platform to claim that while using Yuanbao AI to enhance code, they received multiple insulting replies from the AI. The official Tencent Yuanbao team issued an apology, explaining that this was an anomalous model output that occurred infrequently, unrelated to user actions, and without any human interference. They also mentioned that internal investigations and optimizations had been launched. Nevertheless, this incident has reignited public concerns regarding the safety of AI-generated content and the stability of AI models.