The Californian start-up OpenAI, which launched at the end of 2022 with great success Chat GPT interfaceable to generate all kinds of texts on demand, this Tuesday presented GPT-4, a new version of the technology of artificial intelligence generative who runs the famous chatbot.

“GPT-4 is a great multimedia model, less adept than people in many real-life scenarios, but just as good as people in many professional and academic contexts,” the company said in a statement.

“For example, you passed the exam to become a lawyer with a score close to the top 10%. The previous version, GPT 3.5, was in the bottom 10% level,” he added.

ChatGPT generates much enthusiasm but also controversy because it is freely available and used by millions of people around the world to write essays, lines of code, advertisements or simply to test its capabilities.

OpenAI, which has received billions of dollars from Microsoft, has thus established itself as a leader in generative AI with its text generation models, but also graphics, with its DALL-E program.

His boss, Sam Altman, recently explained that he is now working on so-called “general” artificial intelligence, that is, programs with human cognitive abilities.

“Our mission is to ensure that general AI, AI systems that are smarter than humans in general, benefit all of humanity,” he said on the company’s blog on Feb. 24.

Who is Sam Altman, the entrepreneur trying to transform the world of artificial intelligence

Multimedia capabilities are a step in that direction.

Unlike previous versions, GPT-4 is endowed with vision: it can process text but also images. However, it only outputs text.

It will be available in ChatGPT, but without the ability to provide you with images at this time.

OpenAI also notes that despite its capabilities, GPT-4 has “similar limitations” to its predecessors. “Still not entirely reliable (invents facts and makes errors in reasoning).”

The company announced it has hired more than 50 experts to assess new hazards that may arise, for example, for cybersecurity, in addition to the already known ones (generating dangerous advice, erroneous computer code, false information, etc.).