ChatGPT 4: OpenAI launches improved version of artificial intelligence

A OpenAI launched, this Tuesday (14), long awaited GPT-4, the latest version of the artificial intelligence model. ABOUT GPT-4 will become the brains of a popular chatbot ChatGPT, providing users with enhanced communication options. However, currently only users ChatGPT Plus and developers using the OpenAI API will have access to this exclusive product.

According to the company, the new GPT-4 system promises greater accuracy in text generation, which will provide a more natural and humanized experience with artificial intelligence. The technology is capable of processing up to 25,000 words of text, which allows you to create longer content and search and analyze documents. With these enhanced capabilities, the GPT-4 is expected to become a powerful tool for improving digital communication and making text content creation more efficient.

“We spent six months making GPT-4 more secure and consistent.“, says the OpenAI website.According to our internal estimates, GPT-4 is 82% less likely to respond to banned content requests and 40% more likely to actually respond than GPT-3.5.” concludes.

GPT-4 is the latest addition to the OpenAI language model family, which includes GPT, GPT-2, and GPT-3. The previously released GPT-3 was responsible for the impressive performance of ChatGPT, a chatbot launched in November 2022 that made a big splash in the market with its OCR capabilities.

However, OpenAI acknowledges that the new version of the language model faces AI-specific hurdles: “GPT-4 still has many known limitations that we are working to address, such as social bias, hallucinations, and intruder alerts.“, the site says.We encourage and promote transparency, user education, and broader AI literacy as society embraces these models. We also intend to empower people to contribute to our models.“.

image recognition

One of the main novelties of GPT-4 is the ability to accept commands in image format such as photographs, diagrams, and screenshots, in addition to plain text. This capability allows the system to create image captions and descriptive texts, further expanding its application possibilities.

While this new functionality is a huge achievement for AI technology, OpenAI has clarified that the image recognition feature will not be available initially but will be added at a later date.

To teach the system to recognize and interpret visual commands, OpenAI has partnered with Danish startup Be My Eyes, whose app describes images for blind or visually impaired people. This collaboration represents an important initiative to improve accessibility for people with visual impairments and encourage greater social inclusion.

Source: OpenAI

…..