Addis Ababa, Ethiopia (16 March 2023)
OpenAI has once again raised the bar in the field of natural language processing with the release of its latest model, GPT-4. Released on March 14, 2023, GPT-4 is the most advanced version of the popular chatbot, ChatGPT, and it promises to revolutionize the way we interact with AI.
One of the most exciting features of GPT-4 is its multimodal API, which accepts images as inputs to generate captions, analyses, and reviews. Another impressive aspect of GPT-4 is its ability to analyze the contents of images, a feature that has not yet been released to the public due to concerns over potential misuse. In a recent demo, OpenAI President Greg Brockman showed the power of this feature by feeding GPT-4 a crude pencil sketch of a website he had made in a notebook. In just a few seconds, GPT-4 was able to turn the sketch into working computer code and build a real, functional website.
GPT-4’s capabilities don’t end there. According to OpenAI, the model is more accurate and capable than its predecessors, performing astonishingly well on a variety of tests. For example, GPT-4 scores higher than 90 percent of human test-takers on the Uniform Bar Exam, and it beats 99 percent of humans on the Biology Olympiad. It even aces a number of Advanced Placement exams, including A.P. Art History and A.P. Biology, and gets a score of 1,410 on the SAT.
However, there are cautions and limitations to consider. GPT-4 is not yet available to everyone (only to ChatGPT Plus users), and the team is asking for patience as they roll it out. There is also limited availability of the 32k token model, and a rate limit of 100 messages per 4 hours to GPT-4 in ChatGPT Plus. Additionally, data still cuts off at September 2021, which may limit the model’s ability to generate up-to-date information.
One of the biggest concerns surrounding GPT-4, and AI language models in general, is the potential for disinformation and influenced extremism. While GPT-4’s improved writing capabilities are impressive, they also have the potential to be misused for nefarious purposes.
Despite these caveats, there is no denying that GPT-4 is a major leap forward in the field of natural language processing. With its multimodal API, image analysis capabilities, and improved accuracy and capabilities, GPT-4 is poised to change the way we interact with AI and unlock new possibilities for a wide range of industries.
Source: OpenAI, The New York Times, LinkedIn News