Scholarlink

<p>This article examines how structured prompting frameworks affect the perceived quality of AI-generated social media posts. Using an exploratory study, the RAF, RISE, and CARE frameworks are compared in a&#xa0;product campaign and a&#xa0;microlearning scenario. The basis for this is a&#xa0;theory-based evaluation framework with cognitive, affective, and behavioral dimensions, as well as additional quality and outcome criteria. Six AI-generated posts were evaluated by 17&#xa0;participants. The results indicate context-dependent differences between the frameworks. RAF tends to be more frequently associated with cognitive quality, credibility, and usefulness, while RISE performs better in the campaign context in terms of affect- and intention-related criteria. The evaluation framework can be used as a&#xa0;tool for structured reflection on AI-supported text production in organizations.</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

Prompting-Frameworks und die wahrgenommene Qualität KI-generierter Social-Media-Beiträge.

  • 期刊论文

  • 发表日期：

    2026年6月29日

查看全文
  • Matthias Berg,
  • Michael Ludwig,
  • Sebastian Akinci,
  • Thomas Bleistein

摘要

This article examines how structured prompting frameworks affect the perceived quality of AI-generated social media posts. Using an exploratory study, the RAF, RISE, and CARE frameworks are compared in a product campaign and a microlearning scenario. The basis for this is a theory-based evaluation framework with cognitive, affective, and behavioral dimensions, as well as additional quality and outcome criteria. Six AI-generated posts were evaluated by 17 participants. The results indicate context-dependent differences between the frameworks. RAF tends to be more frequently associated with cognitive quality, credibility, and usefulness, while RISE performs better in the campaign context in terms of affect- and intention-related criteria. The evaluation framework can be used as a tool for structured reflection on AI-supported text production in organizations.