Scholarlink

Background <p>Large language models, such as GPT-4o, have demonstrated strong performance in emotion recognition tasks, yet documented cultural biases raise concerns about cross-cultural validity. Prompt-engineering strategies, including persona prompting and native-language prompting, have been proposed as lightweight approaches to cultural adaptation, but their effectiveness in image-based emotion recognition remains underexamined, particularly in non-Anglophone European contexts. The present study examines whether language choice (Polish vs. English) and persona prompting alter GPT-4o emotional evaluation when benchmarked against a mixed European reference sample from the Nencki Affective Picture System (NAPS).</p> Methods <p>GPT-4o (v.2024-08-06) evaluated 1,276 NAPS images on three emotional dimensions (valence, arousal, approach–avoidance) using a nine-point scale under three conditions: standard GPT-4o, persona-type GPT (Polish university student persona), and Polish-language GPT. Each image received 55 independent evaluations per condition, matching the original NAPS dataset (<i>N</i> = 204 evaluators; 60% Polish nationals, 40% European exchange students). Human–artificial intelligence (AI) agreement was evaluated using Pearson correlations, intraclass correlation coefficients (ICC[2,1]), Bayesian one-way ANOVA, and Anderson–Darling distributional tests.</p> Results <p>All AI conditions positively correlated with human ratings across the three dimensions (<i>r</i> = 0.525–0.809, all <i>p</i> &lt; 0.001). Neither persona nor Polish-language prompting consistently improved agreement over standard GPT-4o. The Polish-language condition aligned more closely with human ratings for arousal but diverged more strongly in approach–avoidance (<i>d</i> = 0.47). Valence showed the strongest correspondence (<i>r</i> = 0.784–0.809; BF₀₁ = 36.55 favoring equivalence), whereas arousal and approach–avoidance exhibited reliable mean-level and distributional divergence. Landscape stimuli produced the highest agreement (<i>r</i> up to 0.933) and faces the lowest (<i>r</i> as low as 0.253). Distributional differences persisted across all nine human–AI comparisons.</p> Conclusions <p>Within the constraints of a mixed European reference sample, persona-based and native-language adaptation strategies did not yield consistent improvements in human–AI agreement. Effects varied across dimensions and analytical methods, indicating that such strategies require dimension-specific empirical validation. AI evaluation aligned acceptably with human ratings for valence and scenic stimuli but diverged for arousal, approach–avoidance, and facial stimuli. These findings underscore the importance of jointly examining rank-order agreement, central tendency, and distributional structure in cross-cultural AI emotion recognition.</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

Comparing human and AI emotional evaluations of images: GPT-4o performance across standard, persona, and language-specific prompting strategies

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月26日

查看全文
  • Maiko Kobayashi,
  • Akira Goto,
  • Akihito Himuro

摘要

Background

Large language models, such as GPT-4o, have demonstrated strong performance in emotion recognition tasks, yet documented cultural biases raise concerns about cross-cultural validity. Prompt-engineering strategies, including persona prompting and native-language prompting, have been proposed as lightweight approaches to cultural adaptation, but their effectiveness in image-based emotion recognition remains underexamined, particularly in non-Anglophone European contexts. The present study examines whether language choice (Polish vs. English) and persona prompting alter GPT-4o emotional evaluation when benchmarked against a mixed European reference sample from the Nencki Affective Picture System (NAPS).

Methods

GPT-4o (v.2024-08-06) evaluated 1,276 NAPS images on three emotional dimensions (valence, arousal, approach–avoidance) using a nine-point scale under three conditions: standard GPT-4o, persona-type GPT (Polish university student persona), and Polish-language GPT. Each image received 55 independent evaluations per condition, matching the original NAPS dataset (N = 204 evaluators; 60% Polish nationals, 40% European exchange students). Human–artificial intelligence (AI) agreement was evaluated using Pearson correlations, intraclass correlation coefficients (ICC[2,1]), Bayesian one-way ANOVA, and Anderson–Darling distributional tests.

Results

All AI conditions positively correlated with human ratings across the three dimensions (r = 0.525–0.809, all p < 0.001). Neither persona nor Polish-language prompting consistently improved agreement over standard GPT-4o. The Polish-language condition aligned more closely with human ratings for arousal but diverged more strongly in approach–avoidance (d = 0.47). Valence showed the strongest correspondence (r = 0.784–0.809; BF₀₁ = 36.55 favoring equivalence), whereas arousal and approach–avoidance exhibited reliable mean-level and distributional divergence. Landscape stimuli produced the highest agreement (r up to 0.933) and faces the lowest (r as low as 0.253). Distributional differences persisted across all nine human–AI comparisons.

Conclusions

Within the constraints of a mixed European reference sample, persona-based and native-language adaptation strategies did not yield consistent improvements in human–AI agreement. Effects varied across dimensions and analytical methods, indicating that such strategies require dimension-specific empirical validation. AI evaluation aligned acceptably with human ratings for valence and scenic stimuli but diverged for arousal, approach–avoidance, and facial stimuli. These findings underscore the importance of jointly examining rank-order agreement, central tendency, and distributional structure in cross-cultural AI emotion recognition.