OpenAI Worried Users May Develop Emotional Connections to AI
Published: August 9, 2024, 11:09 PM
The company behind the AI chatbot ChatGPT is concerned that its realistic software may inadvertently encourage users to form emotional connections with the technology, regardless of their interactions with humans.
“Anthropomorphism involves attributing human traits or characteristics to a non-human entity,” the company stated in a report. “The risk may be increased by the voice features of GPT-4o, which facilitate human-like communications.”
The report coincides with the release of the new voice-enabled version of ChatGPT, GPT-4o. While this allows for more natural interactions, it also raises concerns about users potentially developing misplaced trust in the AI.
OpenAI observed instances among testers where users expressed emotional responses to the AI, including feelings of regret when the interaction ends. The company emphasizes the need for further research to understand the long-term effects of such interactions.
The report also highlights the potential for users to become overly reliant on AI technology, citing concerns about power dependence. This aligns with the growing debate surrounding the responsible use of AI and its impact on human relationships.
OpenAI emphasizes that AI is a tool designed to enhance human lives, but acknowledges the need to address potential risks associated with its development.