Over the past decade, the artificial intelligence (AI) industry, as it relates to the speech and voice recognition industry, has established itself as a multibillion-dollar global market, but at whose expense? In this forum article, we amplify the current critiques of the architectures of large language models being used increasingly in daily life. Our commentary exposes emerging AI accent modification technology and services as agents of racial commodification and linguistic dominance, as it rests on the perceived superiority of standardized US English. We discuss our concern for such services leaching into academia. We argue that this technology follows a standardized language framework, which poses a fundamental problem of being informed by purist monolingual principles. These principles often help to perpetuate and maintain harmful raciolinguistic ideologies that result in language discrimination and the continual framing of the language practices of racially minoritized speakers as deficient. Thus, we write this piece with the intent to expose the fabricated humanity of accent modification technology whose existence perpetuates capitalism’s reliance on dehumanization for economic advancement and the legacy and reproduction of white language superiority.

中文翻译：

---

超越前院：口音改变技术的非人化信息

在过去的十年里，与语音和语音识别行业相关的人工智能（AI）行业已经成为一个价值数十亿美元的全球市场，但谁的利益受损了呢？在这篇论坛文章中，我们放大了当前对日常生活中越来越多地使用的大型语言模型架构的批评。我们的评论揭露了新兴的人工智能口音修改技术和服务是种族商品化和语言主导地位的代理人，因为它依赖于标准化美式英语的感知优势。我们讨论了对此类服务渗入学术界的担忧。我们认为，这项技术遵循标准化的语言框架，这提出了遵循纯粹单语原则的基本问题。这些原则往往有助于延续和维持有害的种族语言意识形态，从而导致语言歧视，并不断将少数种族使用者的语言实践视为有缺陷。因此，我们写这篇文章的目的是揭露口音修改技术捏造的人性，这种技术的存在使资本主义长期依赖非人性化来实现经济进步，以及白人语言优势的遗产和再现。