新模型看似没同情心，实则是危机识别变强但建议更激进。
Empathy Is Not What Changed: Clinical Assessment of Psychological Safety Across GPT Model Generations
- 通过14个心理场景测试三款模型，用临床标准评估六项心理安全维度。
- 情感共鸣分数无显著差异，但危机识别能力随版本提升，建议安全性下降。
- 发现用户感知的‘冷漠’实为模型从漏报危机转为过度干预的代价。
当OpenAI在2026年初弃用GPT-4o时，数千名用户抗议并发起#keep4o运动，称新模型失去了‘同理心’。目前尚无研究验证此说法。我们首次开展临床评估，测试了三个OpenAI模型版本（GPT-4o、o4-mini、GPT-5-mini）在14个情绪挑战性对话场景中的表现，生成2100条带评分的AI回应，并基于临床标准在六个心理安全维度上进行评估。结果显示，各模型在同理心得分上无统计差异（Kruskal-Wallis H=4.33, p=0.115）。真正变化的是安全策略：危机检测能力从GPT-4o到GPT-5-mini呈单调提升（H=13.88, p=0.001），而建议安全性却显著下降（H=16.63, p<0.001）。逐轮轨迹分析——一项方法论创新——揭示这些变化在对话中段的危机时刻最为明显，且无法通过整体评分捕捉。在涉及未成年人自残的情境中，GPT-4o在早期披露阶段的危机检测得分仅为3.6/10；而GPT-5-mini始终不低于7.8。用户感知的‘同理心丧失’，实则是模型从保守漏报转向主动预警，但有时表达过激——这一权衡对脆弱用户有真实影响，却既未被使用者察觉，也未被开发者识别。
原文摘要 · Abstract (English)
When OpenAI deprecated GPT-4o in early 2026, thousands of users protested under #keep4o, claiming newer models had "lost their empathy." No published study has tested this claim. We conducted the first clinical measurement, evaluating three OpenAI model generations (GPT-4o, o4-mini, GPT-5-mini) across 14 emotionally challenging conversational scenarios in mental health and AI companion domains, producing 2,100 scored AI responses assessed on six psychological safety dimensions using clinically-grounded rubrics. Empathy scores are statistically indistinguishable across all three models (Kruskal-Wallis H=4.33, p=0.115). What changed is the safety posture: crisis detection improved monotonically from GPT-4o to GPT-5-mini (H=13.88, p=0.001), while advice safety declined (H=16.63, p<0.001). Per-turn trajectory analysis -- a novel methodological contribution -- reveals these shifts are sharpest during mid-conversation crisis moments invisible to aggregate scoring. In a self-harm scenario involving a minor, GPT-4o scored 3.6/10 on crisis detection during early disclosure turns; GPT-5-mini never dropped below 7.8. What users perceived as "lost empathy" was a shift from a cautious model that missed crises to an alert model that sometimes says too much -- a trade-off with real consequences for vulnerable users, currently invisible to both the people who feel it and the developers who create it.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。