Tada Images/Shutterstock
Generative AI chatbots such as ChatGPT, Gemini, Claude, and others must not be mistaken for actual individuals. They lack authentic personalities, functioning based on an internal system prompt that is not visible to the user, which governs their actions. Nevertheless, the companies developing these AIs consistently enhance their offerings, which includes modifying their personalities. Recently, both ChatGPT (GPT-5) and Claude received notable updates that are expected to affect the user interaction.
While OpenAI made adjustments to the GPT-5 personality to counteract complaints that arose shortly after the introduction of this advanced ChatGPT version, Anthropic equipped Claude with the capability to autonomously terminate conversations if the AI assesses that a particular interaction could be detrimental. These personality modifications are expected to be present in the ChatGPT and Claude applications, although not all users may observe them immediately. The changes to GPT-5 are subtle. In the case of Claude, the chatbot will only conclude conversations in exceptional circumstances.