“OpenAI's current research direction involves integrating multiple modalities, such as text, images, voice, and video, into its large-scale models.”