OpenAI has launched its latest artificial intelligence model, GPT-4o, with several updates including a new desktop service, vision reasoning, and enhanced voice assistant features. Unlike previous versions, GPT-4o will be available for free, a departure from the subscription model for previous ChatGPT 4 users.

Speculation about OpenAI launching a ChatGPT-powered search competitor to Google has been put on hold. CEO Sam Altman confirmed that this announcement was not related to search.

When Chief Technology Officer Mira Murati introduced the model, she highlighted its potential to revolutionise human-machine interaction.

According to Tim Sharp, founder and Principal of Gen8, a generative AI strategy firm, his first impression of the updated platform was “Fast. Very fast.”

“Zipped through my usual demos, including some pretty complex data analysis prompts, from large spreadsheets. That’s big for me.”

He described ChatGPT 4o as technically impressive, especially having one (‘native’, ‘multimodal’) neural network, instead of multiple models strung together.  “I have no idea how they’ve done this, but looking forward to learning more.”

However, he also said the ultra-lifelike voice assistants sit uncomfortably with him. “I really believe we should be humanising AI less, not more. Did Spike Jonze ever really intend ‘Her’ to feel this close to reality (or maybe that was the exact prediction?)”

“Ditto for making GPT4 ‘free’. This is exciting, but less when you think about the tradeoffs. Why keep paying for Plus? I haven’t read into this in detail but one of the big implications is privacy. If it’s free, we are the product?”

Updates

The updates to the model include improved international language capabilities, the ability to upload and analyse images, audio, and text documents, and new voice functionalities.

Additional reporting by Andrew Birmingham