Scholarlink

<p>Biomedical engineers produce knowledge and artifacts. Across the life cycle of an idea, errors can creep in. In this letter, we propose the term linguistic, orthographic, and typographical errors in science (LOTS) to represent a category of errors that threaten the truthfulness and integrity of scientific literature and engineering projects. They include documented cases of the misuse of generative artificial intelligence (GAI). LOTS consist of four categories: (1) simple spelling errors; (2) the semantic deformation of technical terms, in the form of ‘tortured phrases’; (3) letter or symbol-switching; and (4) formatting errors that impact the veracity of knowledge, or distort the precision of scientific representation, such as the absence or overuse of capitalization, the incorrect use or absence of italicization, the failure to deanonymize information, cloned template text, or GAI-generated “hallucinations.” We introduce small analyses to assess the incidence of dopamine-ß-hydroxylase, ß-secretase, and ß-adrenoreceptors (erroneous <i>Eszett</i> formats) supposedly representing dopamine-β-hydroxylase (DBH), β-secretase, and β-adrenoreceptors, respectively, in PubMed. We suggest that virtuous biomedical engineers should address LOTS. To improve the screening of LOTS, we argue in favor of a common framework for scholarly text integrity analysis.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Linguistic, Orthographic, and Typographical Errors in Science (LOTS): A Growing Threat to Academic Integrity

  • 期刊论文

  • 发表日期：

    2026年3月21日

查看全文
  • Jaime A. Teixeira da Silva,
  • Timothy Daly

摘要

Biomedical engineers produce knowledge and artifacts. Across the life cycle of an idea, errors can creep in. In this letter, we propose the term linguistic, orthographic, and typographical errors in science (LOTS) to represent a category of errors that threaten the truthfulness and integrity of scientific literature and engineering projects. They include documented cases of the misuse of generative artificial intelligence (GAI). LOTS consist of four categories: (1) simple spelling errors; (2) the semantic deformation of technical terms, in the form of ‘tortured phrases’; (3) letter or symbol-switching; and (4) formatting errors that impact the veracity of knowledge, or distort the precision of scientific representation, such as the absence or overuse of capitalization, the incorrect use or absence of italicization, the failure to deanonymize information, cloned template text, or GAI-generated “hallucinations.” We introduce small analyses to assess the incidence of dopamine-ß-hydroxylase, ß-secretase, and ß-adrenoreceptors (erroneous Eszett formats) supposedly representing dopamine-β-hydroxylase (DBH), β-secretase, and β-adrenoreceptors, respectively, in PubMed. We suggest that virtuous biomedical engineers should address LOTS. To improve the screening of LOTS, we argue in favor of a common framework for scholarly text integrity analysis.