Scholarlink

<p>Educators and designers alike incorporate images in teaching and learning to create shared meaning, and to engage learners. With the growth of generative artificial intelligence (genAI) incorporating images is as easy as typing a quick prompt. Think of an idea, type a prompt, and an image synthesizer can create it. Image synthesizers are genAI programs that can generate an image from a text prompt. With the growth of genAI, a common problem among educators of finding an appropriate or desired image or illustration to support instruction may be solved by utilizing an image synthesizer. Currently, though, the relevance and accuracy of these types of images are relatively unknown. Therefore, a systematic review of AI-generated images was warranted. In this novel study, we conducted a visual content analysis of AI generated images (<i>N</i> = 291)&#xa0;representing STEM-related occupations (e.g., scientist, mathematician). The analysis of these AI-generated images from three AI image generation programs indicated gender bias (21.3%), racial bias (18.9%), eyes disfigurement (17.9%), and hand inaccuracies (22.3%). A four-step decision-making inclusion process is provided for practitioners to raise awareness when including AI-generated images in instruction.</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

Generative Artificial Intelligence Images for Instruction: A Content Analysis

  • 期刊论文

  • 发表日期：

    2025年9月6日

查看全文
  • Laurie O. Campbell,
  • Glenn W. Lambie,
  • B. Grant Hayes,
  • Richard Hartshorne

摘要

Educators and designers alike incorporate images in teaching and learning to create shared meaning, and to engage learners. With the growth of generative artificial intelligence (genAI) incorporating images is as easy as typing a quick prompt. Think of an idea, type a prompt, and an image synthesizer can create it. Image synthesizers are genAI programs that can generate an image from a text prompt. With the growth of genAI, a common problem among educators of finding an appropriate or desired image or illustration to support instruction may be solved by utilizing an image synthesizer. Currently, though, the relevance and accuracy of these types of images are relatively unknown. Therefore, a systematic review of AI-generated images was warranted. In this novel study, we conducted a visual content analysis of AI generated images (N = 291) representing STEM-related occupations (e.g., scientist, mathematician). The analysis of these AI-generated images from three AI image generation programs indicated gender bias (21.3%), racial bias (18.9%), eyes disfigurement (17.9%), and hand inaccuracies (22.3%). A four-step decision-making inclusion process is provided for practitioners to raise awareness when including AI-generated images in instruction.