San Francisco: OpenAI on Monday released a higher and more human-like version of its artificial intelligence technology that powers the popular creativity tool ChatGPT and made it free to all users.
The update to OpenAI’s flagship product comes days before the search engine giant, which competes with Google’s ChatGPT leader, unexpectedly announced its own Gemini AI tool.
A new model of the GPT-4o — the “O” stands for omni — will be available in the OpenAI product in the next few weeks, he said, adding that paying customers will have unlimited access to the tool.
The company says the model can generate content or understand commands from voice, text or images.
“The new audio (and video) mode is the best computer interface I’ve ever used. It feels like the AI from the movies,” said OpenAI CEO Sam Altman in a blog post.
Altman has previously cited Scarlett Johansson’s character in Her as inspiration for where he wants his relationship with AI to go.
Engineers from Myrati and OpenAI demonstrated the new capabilities of GPT-4o in a virtual event, challenging the chatbot replication version of ChatGPT.
In the demo, most of the OpenAI staff asked questions to the ChatGPT voice, answering with jokes and banter like humans.
The bot acts as an English-to-Italian translator, interprets facial expressions, and solves users’ difficult algebra problems.
The company claims that GPT-4o has the same advantages as its predecessor when it comes to text, conversation and encoding intelligence, and sets new industry standards for multi-lingual conversation, voice and vision.
In a demonstration, ChatGPT successfully explained the surroundings of an employee through a smartphone camera, speaking in a friendly feminine voice, not unlike the AI bot in the movie Her.
