Tech

OpenAI’s GPT-4 ‘Show performance at a human level’, will respond to both image and text input


The company behind the ChatGPT app that writes essays, poems or computer code on command released Tuesday a long-awaited update of its artificial intelligence (AI) technology, it said. will be safer and more accurate than its predecessor.

GPT-4 has been widely awaited since then ChatGPT exploded in late November, wowing users with capabilities based on an older version of OpenAI technology, known as the big language model.

“We created GPT-4the latest milestone in OpenAI’s efforts to scale deep learning,” a company blog said, adding that the AI ​​technology “reveals human-level performance” in a number of specialist tasks. and academic.

The company says the model is “more innovative and collaborative than ever” and will “solve tough problems with greater precision” than previous versions.

With this update, the text responses from GPT-4 will be more accurate and — in the future — will come from both image and text inputs in a huge leap forward in technology, though This aspect has yet to be released.

For example, if a user submits pictures of the inside of a refrigerator, the GPT-4 will not only identify exactly what is there, but also concoct what can be prepared with those ingredients.

openAI said that it is working with a partner company, Be My Eyes, to prepare for the next step.

Much of the new model’s firepower is now available to the public through ChatGPT PlusOpenAI’s paid subscription and on Microsoft’s AI-powered version Bing Search engine is currently being tested.

OpenAI is powered by Microsoftwhich earlier this year said it would fund the research firm with billions of dollars.

The Windows maker then quickly integrated the technology into its Bing search engine, The edge browsers and other products.

Microsoft’s active adoption of ChatGPT has sparked a race against Google announced its own versions of AI technology, with Amazon, Baidu And meta also waded in, eager not to be left behind.

Less ‘hallucination’

OpenAI says that the new version is much less likely to get sidetracked than its previous chatbot with its widely reported interactions with ChatGPT or Bing’s chatbot in which users were presented with lies, insults or other so-called “hallucinations”.

“We spent six months making GPT-4 safer and more relevant. GPT-4 is 82% less likely to respond to requests for unauthorized content and more likely to respond to requests. 40% more realistic recovery,” OpenAI said.

Founder Sam Altman admits that despite expectations, the GPT-4 “is still flawed, still limited, and it still looks more impressive on first use than it does after you’ve spent more time with it.”

Robert Vesoul, CEO of Illuin Technology, a French AI startup, said: “The power of algorithms will increase, but it will not be the second revolution.

Vesoul questions the safety measures taken by OpenAI, which has been criticized by the billionaire Elon Musk that the company is overly limiting its AI speech to avoid more embarrassing reactions.

“I’m not sure if I want the AI ​​to block responses on unknown topics… Should the AI ​​decide if I smoke?” Vesoul told AFP.

Other companies collaborating on GPT-4 implementations including Morgan Stanley will use AI to help guide bankers and their customers.

“Basically, you have the knowledge of the best in Wealth Management—instantly,” Morgan Stanley Jeff McMillan said in a statement.

Other partners include Khan Academy, the online tutoring giant, and Stripe, a financial app that will use GPT-4 for fraud prevention and other uses.


Affiliate links can be generated automatically – see ours Moral standards for details.

news7f

News7F: Update the world's latest breaking news online of the day, breaking news, politics, society today, international mainstream news .Updated news 24/7: Entertainment, Sports...at the World everyday world. Hot news, images, video clips that are updated quickly and reliably

Related Articles

Back to top button