Sam Altman Says OpenAI Doesn’t Work on GPT-5 – 04/14/2023 – Tech

OpenAI is not working on the development of GPT-5, a possible more powerful version of the artificial intelligence (AI) behind ChatGPT (a tool capable of generating texts), said the startup’s chief executive, Sam Altman, at an event held on Thursday. fair (13) at MIT (Massachusetts Institute of Technology), in the United States.

Altman addressed the topic when asked by MIT researcher Lex Friedman about the letter that asked for a pause in the development of generative AIs (capable of producing texts, images or other works) and gathered signatures from Elon Musk, writer Yuval Harari and co-founder from Apple Steve Wozniak,.

The document was released at the end of March by Future of Life, an institute advised and funded by Musk. It was a response to rumors that GPT-5 could achieve artificial general intelligence, when the technology could perform any intellectual task possible for humans.

In the same month, OpenAI had released GPT-4, a new technology that, on the one hand, impressed the public with more accurate responses than ChatGPT (equipped with GPT-3.5), but also raised concerns about new risks of misinformation, cybersecurity and the deterioration of the job market.

Altman claims to share the concern about problems mentioned in the letter signed by Musk, but considers that the document ignores technical nuances. “It doesn’t discuss where in the process we need to pause.”

“It is increasingly necessary to add security as AI models gain new capabilities,” adds Altman. According to the executive, OpenAI prioritizes the improvement of GPT-4, which underwent security training for six months before being presented to the public.

OpenAI, however, will continue with the strategy of releasing imperfect products to the public, according to Altman. ChatGPT, for example, delivers inaccurate or wrong answers. Pioneering AI companies like Google and Microsoft were walking on eggshells about the technology after gaffes like the 2016 chatbot Tay, which reproduced hate speech on Twitter.

“Our goal is to get the world involved in the discussion about artificial intelligence, think about it and gradually improve and build new institutions or adapt them to understand the future that society wants”, says the executive, who heads the business part of OpenAI .

The non-profit arm of the startup, chaired by Greg Brockman, says it has a mission to “develop artificial general intelligence that benefits all of society”.

Altman rebuilt OpenAI as what he called a limited-profit company in 2019. That allowed him to seek billions of dollars in funding, promising profit from investors like Microsoft. But those profits are limited, and any additional revenue will be pumped back into the non-profit organization OpenAI, founded in 2015.

The executive’s idea is that OpenAI will capture much of the world’s wealth through the creation of AGI and then redistribute that wealth to the people. In Napa, in conversation with the New York Times, he cited various numbers – $100 billion, $1 trillion, $100 trillion.

In a 2021 manifesto, Altman claims that artificial general intelligence has the potential to transform the economy and replace expensive professionals such as higher education professors and doctors. Therefore, he calls on governments to discuss the impacts of technology since then.