OpenAI has revealed ChatGPT-4o (the ‘o’ is for ‘omni’) promising a more conversational feel. The company described GPT-4o as a “step toward a much more natural human-computer interaction.” The tool has a lot of new features and also communicates much more naturally and with emotions.
Because GPT-4o is much better and much faster than OpenAI’s previous models, the tool also has some notable new features. One of these is the ability to translate conversations in real time.
Before starting a conversation, users can ask GPT-4o to leave that conversation. You can do that, for example, by saying “translate everything in Spanish to English and everything in English to Spanish.” During the announcement of this update, OpenAI CTO Mira Murati immediately demonstrated how that works. In doing so, he immediately noticed how quickly and naturally this translation happens.
In addition to a Spanish demonstration, OpenAI also held an Italian-English demo online, which worked just as well. In total, the tool is available in more than 50 languages.
It is unclear when this feature will be available to the general public. However, it is already clear that it will revolutionise international communication. It has long been predicted that AI will solve language barriers in the future.
According to OpenAI, GPT-4o is significantly better on Turbo performance on text in English and code, with significant improvement on text in non-English languages, while also being faster and cheaper in the API. GPT-4o is especially better at vision and audio understanding.
The revamped Voice Mode GPT-4o trades in a multi-model system used in versions 3.5 and 4 new model end-to-end across text, vision, and audio, meaning that all inputs and outputs are processed by the same neural network. Because GPT-4o is our first model combining all of these modalities, we are still just scratching the surface of exploring what the model can do and its limitations.
Perhaps as a response to concerns over the rapid pace of AI development OpenAI emphasised it plans to progress in a way that will preserve accuracy and provide a safe user experience.
“GPT-4o has safety built-in by design across modalities, through techniques such as filtering training data and refining the model’s behaviour through post-training. We have also created new safety systems to provide guardrails on voice outputs,” the company wrote on its website.
“We’ve evaluated GPT-4o according to our Preparedness Framework and in line with our voluntary commitments. Our evaluations of cybersecurity, CBRN, persuasion, and model autonomy show that GPT-4o does not score above Medium risk in any of these categories. This assessment involved running a suite of automated and human evaluations throughout the model training process. We tested both pre-safety-mitigation and post-safety-mitigation versions of the model, using custom fine-tuning and prompts, to better elicit model capabilities.”
GPT-4o is available in the free tier, and to Plus users with up to 5x higher message limits. A new version of Voice Mode with GPT-4o within ChatGPT Plus will be released in the coming weeks.
TechCentral Reporters
Subscribers 0
Fans 0
Followers 0
Followers