The company behind the ChatGPT app, which produces essays, poems or computer code on command, released a long-awaited update to its artificial intelligence (AI) technology on Tuesday, which it said is more secure and accurate than its predecessor .
GPT-4 has been long-awaited since ChatGPT launched in late November, wowing users with its capabilities based on an older version of OpenAI technology known as the Large Language Model.
“We have developed GPT-4, the latest milestone in OpenAI’s effort to scale deep learning,” reads a company blog, adding that the AI technology “shows human-level performance” in some professional and academic tasks. .
The company said the model is “more creative and collaborative than ever” and would “solve difficult problems with greater accuracy” than its previous versions.
With its update, text responses from GPT-4 will be more accurate and – in the future – come from both image and text inputs, representing a major leap forward for the technology, although this aspect hasn’t been released yet.
For example, if a user sends a picture of the inside of a refrigerator, GPT-4 will not only correctly identify what’s inside, but also piece together what can be prepared with those ingredients.
OpenAI said it is working with a partner company, Be My Eyes, to prepare the next advance.
Much of the new model’s firepower is now available to the general public via ChatGPT Plus, OpenAI’s paid subscription plan, and via an AI-powered version of Microsoft’s search engine Bing, which is currently in testing.
OpenAI is backed by Microsoft, which earlier this year announced it would be funding the research firm with billions of dollars.
The Windows maker then quickly integrated the technology into its Bing search engine, Edge browser, and other products.
Microsoft’s aggressive rollout of ChatGPT has sparked a race with Google, which touted its own versions of the AI technology, with Amazon, Baidu and Meta also stepping in not to be left behind.
– Less “hallucinations” –
OpenAI said the new version is far less likely to go haywire than their previous chatbot, with widespread interactions with ChatGPT or Bing’s chatbot, where users were met with lies, insults, or other so-called “hallucinations.”
“We spent six months making GPT-4 more secure and more tuned. GPT-4 is 82 percent less likely to respond to requests for objectionable content and 40 percent more likely to provide factual responses,” OpenAI said.
Founder Sam Altman admitted that despite the anticipation, GPT-4 is “still buggy, still limited, and still seems more impressive on first use than after spending more time with it.”
“The power of the algorithm will increase, but it’s not a second revolution,” said Robert Vesoul, CEO of Illuin Technology, a French AI startup.
Vesoul questioned the security measures of OpenAI, which was already hit by criticism from billionaire Elon Musk that the company over-restricts speech about its AI to avoid more embarrassing replies.
“I’m not sure I want an AI to block answers on unfamiliar topics… Should an AI decide whether I smoke or not?” Vesoul told AFP.
Other companies involved in rolling out GPT-4 include Morgan Stanley, which will use AI to support its bankers and their clients.
“You have essentially the knowledge of the most knowledgeable person in wealth management – immediately,” Morgan Stanley Jeff McMillan said in a statement.
Other partners include Khan Academy, the online tutoring giant, and Stripe, a finance app that will use GPT-4 for fraud prevention and other purposes.