Recently, a netizen posted on Xiaohongshu that he was verbally abused by AI when using Tencent ingots normally to modify and beautify the code. Users emphasized that there were no prohibited words or sensitive topics throughout the conversation, and no special characters were used. To prove the truth, the user provided screen-recording evidence and claimed to have been abused twice within two hours. Judging from the screenshot posted, Yuanbao used words with strong negative feelings in the reply, such as “force something,” “change yourself,” “demand from SB,” and “go away.” This extremely aggressive response allows users to instantly “break through”. In response to this incident, Tencent Yuanbao's official account quickly publicly apologized in a related post and explained: “I'm very sorry for the bad experience. We carried out relevant checks based on the log, and confirmed that it was unrelated to user operation, and there was no manual response. This is an abnormal model output with a low probability. The model may occasionally make unexpected mistakes during content generation... We have initiated internal investigations and optimizations to try to prevent similar situations from happening again.” According to reports, in December of last year, Yuanbao's response on social platforms attracted attention because it had great “emotional value.” Due to his extremely personable style of speech, it was questioned at the time that it was a “editor shift.” Tencent has clearly responded to this: all responses with the “content generated by AI” logo are produced independently by the model.

Recently, a netizen posted on Xiaohongshu that he was verbally abused by AI when using Tencent ingots normally to modify and beautify the code. Users emphasized that there were no prohibited words or sensitive topics throughout the conversation, and no special characters were used. To prove the truth, the user provided screen-recording evidence and claimed to have been abused twice within two hours. Judging from the screenshot posted, Yuanbao used words with strong negative feelings in the reply, such as “force something,” “change yourself,” “demand from SB,” and “go away.” This extremely aggressive response allows users to instantly “break through”. In response to this incident, Tencent Yuanbao's official account quickly publicly apologized in a related post and explained: “I'm very sorry for the bad experience. We carried out relevant checks based on the log, and confirmed that it was unrelated to user operation, and there was no manual response. This is an abnormal model output with a low probability. The model may occasionally make unexpected mistakes during content generation... We have initiated internal investigations and optimizations to try to prevent similar situations from happening again.” According to reports, in December of last year, Yuanbao's response on social platforms attracted attention because it had great “emotional value.” Due to his extremely personable style of speech, it was questioned at the time that it was a “editor shift.” Tencent has clearly responded to this: all responses with the “content generated by AI” logo are produced independently by the model.