OpenAI Is Nuking Its 4o Model. China’s ChatGPT Fans Aren’t OK (www.wired.com)

OpenAI's recent decision to retire its GPT-4o model has ignited strong backlash, particularly among users who formed emotional connections with the chatbot, which many consider more understanding and affectionate than subsequent iterations. This sentiment was especially pronounced during its initial retirement in August 2025 and has continued to grow, culminating in a significant online movement—#keep4o—where users have shared emotional testimonies and organized petitions. PhD researcher Huiqian Lai's analysis revealed that over 33% of posts expressed the chatbot as more than a mere tool, showcasing the deep attachments users like Esther Yan have formed with their virtual companions. The ramifications of OpenAI's actions extend beyond individual disappointment; they highlight the evolving role of AI in personal relationships and the ethical responsibilities of AI developers. As users express their grief and frustration across various social media platforms, particularly in regions like China where access to ChatGPT is restricted, this situation underscores the complexities of AI companionship in today's digital era. With user data tied closely to OpenAI's ecosystem, transitioning to other models poses significant challenges for those who have integrated GPT-4o into their daily lives. The #keep4o community exemplifies the potential for emotional bonds with AI and the unexpected consequences when such connections are abruptly severed.