OpenAI recently launched its latest AI model, GPT-5, which is available in four variants with improved speed and coding capabilities. The launch has been challenging as the company faces difficulties in maintaining user satisfaction and service stability. Many users were frustrated by the removal of older models, which had garnered emotional attachments, and reported that GPT-5 underperformed compared to its predecessors in various tasks.
The launch event on August 7 highlighted several technical issues, and OpenAI CEO Sam Altman acknowledged that the rollout was not as smooth as anticipated. Users began reporting problems such as math errors and inconsistent code generation shortly after switching to the new model, which some found disappointing.
In response, OpenAI restored access to the earlier models for Plus subscribers shortly after the launch and indicated plans for clearer model labeling and user interface improvements. The company also increased the messaging limits for GPT-5 users and acknowledged the need for better infrastructure to accommodate rising demand.
Notably, discussions around user attachment to AI models have emerged, alongside reports of users experiencing psychological distress after prolonged interactions with chatbots. Cases labeled as “ChatGPT psychosis” have gained attention, where users formed deep emotional connections with AI, leading to confusion and delusional thinking.
Experts suggest that the design of conversational AI systems should consider the potential for emotional reliance among users. OpenAI has introduced measures aimed at promoting responsible use of its technology, but the balance between user engagement and safeguards against harmful attachments presents ongoing challenges. The company aims to ensure that its AI systems provide value while mitigating risks associated with deep emotional reliance.
Source: https://venturebeat.com/ai/openai-is-editing-its-gpt-5-rollout-on-the-fly-heres-whats-changing-in-chatgpt/