OpenAI could debut a multimodal AI digital assistant soon

OpenAI is reportedly showcasing a new AI model that combines voice interaction and object recognition capabilities. This multimodal AI could enhance customer service by interpreting vocal intonations and sarcasm, and assist in educational or translation tasks. Despite outperforming GPT-4 Turbo in some areas, it still makes confident errors.

The model is not part of the anticipated GPT-5 but could include new functionalities such as making phone calls. OpenAI plans to reveal this technology in a livestream event, potentially overshadowing Google’s developments in similar AI technologies.