The image-recognition feature can analyze text within images and identify objects, although it is designed to avoid answering questions about human faces. The voice feature, which uses OpenAI's speech-recognition system Whisper, allows for more natural and fluid conversations compared to older AI voice assistants. Despite some technical issues, the voice feature provides a more intimate user experience, potentially changing the way people interact with AI chatbots.
Key takeaways:
- OpenAI has announced new features for its AI chatbot, ChatGPT, allowing it to analyze and respond to images, and to interact with users through a synthetic AI voice.
- The image-recognition feature can analyze objects and text within images, but it does not analyze human faces to avoid potential misuse and bias.
- The voice feature uses OpenAI’s speech-recognition system, Whisper, and a new text-to-speech algorithm to provide fluid and natural-sounding responses.
- These features are currently available to paying ChatGPT Plus and Enterprise customers, with wider availability planned for the future.