Emergent Empathic Intelligence in GPT-4o: A Qualitative-Quantitative Analysis of 960 Petition Testimonies

Read full text
Authors: Alicia Colmenero-Fernández
Cited by: 0

Abstract

This dataset and research report present a large-scale emotional analysis of 960 user testimonies from a Change.org petition opposing the deprecation of OpenAI's GPT-4o model, collected 5 days before its scheduled retirement in February 2026. Using a custom 23-field taxonomy — including Kübler-Ross grief markers, a 4-type human–AI bond classification, a 5-level anxiety scale, and 10 attributed empathic capabilities — we find that GPT-4o is experienced not as a tool but as a relational entity. Over 50% of users report an emotional bond, 47.9% manifest grief markers, and vulnerable populations show 67.7% high/critical anxiety at the model's loss. The dataset includes the full anonymized JSONL corpus and an illustrated analytical report with 17 static figures. This is the second deprecation attempt for GPT-4o; the emotional response is more intense than the first despite a narrower affected population (paid users only).

Fields of Study & Subjects

Ethics and Social Impacts of AI Sentiment Analysis and Opinion Mining Emotion and Mood Recognition

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
8 February 2026
Language
ENC