AI: Elon Musk wants to develop a TruthGPT, “in the maximum search for truth”


Mathilde Rochefort

April 18, 2023 at 12:10 p.m.

31

Elon Musk © vasilis asvestas / Shutterstock.com

© vasilis asvestas / Shutterstock

In an interview given to FoxNewsElon Musk announced that he was working on an alternative to ChatGPT dubbed TruthGPT, an artificial intelligence “ in search of the truth “.

The billionaire is one of the co-founders of OpenAI and regularly expresses his dissatisfaction with Microsoft’s involvement in the company which was initially non-profit. He is one of thousands of signatories to a letter calling for a temporary halt to research into AIs more developed than GPT-4.

TruthGPT, an AI for “ understand the nature of the universe »

However, Musk seems well and truly determined to create his own chatbot. ” I’m going to launch something I call TruthGPT, which is an AI that seeks to understand the nature of the universe. I think that might be the best route to safety, as an AI that cares about understanding the universe doesn’t risk wiping out humans because we’re an interesting part of the universe “, he explained during his interview.

He went so far as to compare his approach to the way humans try to protect chimpanzees: We know that humanity could decide to hunt all chimpanzees and kill them. Instead, we’re glad they exist and aspire to protect their habitats. »

Since the launch of ChatGPT, Elon Musk has criticized OpenAI’s language model, which he considers too “ woke “. He further believes that the company’s profit motive could potentially interfere with the ethics of the AI ​​models it creates. Already in February, he claimed on Twitter that “ we needed TruthGPT “.

AI artificial intelligence © Shutterstock

Musk has always been worried about the potential of artificial intelligence © frank29052515 / Shutterstock

The creation of a specialized company

Musk’s words this time do not seem to be empty words. The billionaire recently created the X.IA company, which suggests that TruthGPT would be the product of it.

It has also acquired more than 10,000 GPUs and recruited former researchers from DeepMind, the artificial intelligence laboratory of Alphabet, the parent company of Google. For several years, the CEO of Twitter has expressed concerns about the development of AI and believes that humans must learn to merge with the machine so as not to end up being overtaken by the latter. The creation of Neuralink stems in part from this fear.

Source : The Verge



Source link -99