The fourth version of the revolutionary artificial intelligence chatbot is known as GPT-4. This technology will teach machines to understand natural language and is expected to introduce new features including video processing and multimodality.

For a language model to be multimodal, it must be able to interpret both textual and visual data. The current version of ChatGPT can only generate text responses. However, with the development of GPT-4, it will be possible to generate not only text, but also images and videos. The upcoming version of GPT-4 is expected to be a significant advance over GPT-3.5, boasting a larger context window, higher reliability, and new development features.

The multimodal language functions of the GPT-4 system offer the additional advantage of managing input and output data in different languages. This feature allows users to enter requests in one language, such as Italian, while generating responses in another language, such as English. Such an innovation could also be extremely beneficial for translation tools.

In order to commercialise OpenAI technology, Microsoft has acquired a 49% stake for $10 billion and is integrating AI technology into its products such as Teams and the Bing search engine. As the industry continues to develop, further advances in LLM (large language models) and multimodal models can be expected, bringing new opportunities and possibilities. The imminent arrival of GPT-4 is likely to generate considerable interest and could represent a significant milestone in the development of AI technology.