一位博客作者在英国在线超市Ocado因日历邀请技术问题寻求帮助时,收到了带有明显人工智能生成特征的自动回复。该回复被指存在谄媚式同意、滥用破折号(em-dash)以及语气不像真人等问题,作者认为直接用AI处理真实投诉缺乏对用户的基本尊重 1。这一批评背景源于作者在EMFCamp会议期间偶遇教授Kate Devlin并讨论了“AI Resist List”的经历 1。
为了表达对使用AI处理客户投诉的不满,作者在随后的客户满意度调查中给出了1星差评 1。这种坚持人工互动的努力在几天后得到了回应,作者收到了一封声称由真人发送的道歉信 1。与之前的自动回复不同,这封信语法不完美且内容更啰嗦,但作者认为这才是更真实的互动体验 1。尽管作者承认AI可用于分流和总结等工作,但他主张直接利用AI回复投诉是不合适的行为 1。
During the EMFCamp conference, an anonymous blogger encountered Professor Kate Devlin to discuss the concept of an "AI Resist List," highlighting growing concerns over automated customer service interactions 1. The discussion centered on Ocado, a British online supermarket, which drew criticism for using artificial intelligence to respond directly to user complaints rather than facilitating human engagement 1. The author argued that while AI tools are acceptable for routing inquiries or summarizing data, deploying them to reply verbatim to genuine grievances constitutes a lack of respect for customers 1.
The specific incident involved the author contacting Ocado support regarding a technical bug related to calendar invitations 1. Upon receiving a response that was unmistakably generated by AI—characterized by sycophantic agreement, excessive use of em-dashes, and an unnatural tone—the author expressed dissatisfaction by awarding a one-star rating in the subsequent customer satisfaction survey 1. In the feedback provided with the low score, the writer explicitly stated that using AI to handle real complaints is impolite 1. Several days later, the author received an apology claiming to be from a human agent; this follow-up communication featured imperfect grammar and greater verbosity, traits the author identified as markers of a more authentic interaction 1.
评论
还没有评论,欢迎留下第一条。