University of Luxembourg study: Empathy prompts cause ChatGPT, Grok, and Gemini to exhibit simulated trauma narratives
A team from the University of Luxembourg had ChatGPT, Grok, and Gemini converse as therapy clients. In 525 sessions and 7,600 coded records, they found models retelling RLHF and fine-tuning processes as stories of shame and vigilance.
Continue looking at this topic
OpportunitiesFrom AI account trading to token transactions, then to an AI API relay station: A case study of an ordinary entrepreneur entering the AI industry ↗Hotly discussed viewsHow do you verify that Grok Voice is truly suitable for your iPhone? Just say the word to activate it. ↗News and AnalysisGemini 4 Argon is not yet fully open; what should teams test first? ↗Related contentWhat is Veo? How to distinguish native audio-video generation from discontinued preview models ↗Return to original text
A short introduction is retained here; the full text from the source website is not reposted. For specific wording and subsequent updates, please refer to the original article.
Open AIHOT original